WP44: Part V (Agent E, 98 refs), Part IX results (21 tables), taxonomy validation metrics, KG extraction (11,086 works, 6.16 USD), A-008; work paused by the PI 2026-11-03 ~21:00

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

Simon-Pierre Boucher committed 5 h ago (Nov 4, 2026) parent 10f8005 Browse files

8 changed files +6,054 −0

added data/metadata/extraction_usage_claude-sonnet-5-5.json +6 −0

++{
++ "model": "claude-sonnet-5-5",
++ "calls": 275,
++ "cost_usd": 6.16,
++ "works_done": 1460
++}
+ \ No newline at end of file

modified protocol/amendments.csv +1 −0

				`@@ -7,3 +7,4 @@ A-004,2026-11-03,3,"Knowledge-graph extraction from abstracts of the high-priori`
7	A-005,2026-11-03,1,"Exclusion code E7 redefined as the generic-deposit exclusion (A-002); the CS penumbra is a membership flag, not an exclusion code","Part IV writer found the protocol and code labels inconsistent",executing agent (pending PI)	7	A-005,2026-11-03,1,"Exclusion code E7 redefined as the generic-deposit exclusion (A-002); the CS penumbra is a membership flag, not an exclusion code","Part IV writer found the protocol and code labels inconsistent",executing agent (pending PI)
8	A-006,2026-11-03,1,"CS-core rule: an arXiv primary category cs.* is a sufficient membership signal (previously required a CS topic as well, which excluded the arXiv-linked works recovered by the second pass); PRISMA partition core + penumbra + neither asserted","Agent K found 6,263 post-exclusion records in neither set and the clause unreachable for second-pass works",executing agent (pending PI)	8	A-006,2026-11-03,1,"CS-core rule: an arXiv primary category cs.* is a sufficient membership signal (previously required a CS topic as well, which excluded the arXiv-linked works recovered by the second pass); PRISMA partition core + penumbra + neither asserted","Agent K found 6,263 post-exclusion records in neither set and the clause unreachable for second-pass works",executing agent (pending PI)
9	A-007,2026-11-03,1,"The OpenAlex API increment (works published 2026-10-01 → 2026-11-02 fetched after the snapshot; raw pages archived, 64/65 day-slices, 1 truncated) is NOT merged into the analytical corpus of version 1.0; the effective end of the corpus is the snapshot date 2026-10-14 and the 2026 series are flagged incomplete","merging would require a fourth corpus build after classification had started; the increment is kept for the next version and listed as a limitation",executing agent (pending PI)	9	A-007,2026-11-03,1,"The OpenAlex API increment (works published 2026-10-01 → 2026-11-02 fetched after the snapshot; raw pages archived, 64/65 day-slices, 1 truncated) is NOT merged into the analytical corpus of version 1.0; the effective end of the corpus is the snapshot date 2026-10-14 and the 2026 series are flagged incomplete","merging would require a fourth corpus build after classification had started; the increment is kept for the next version and listed as a limitation",executing agent (pending PI)
		10	+A-008,2026-11-03,4,"pivot_table without aggfunc averaged the (year, period) rows of category_counts_yearly in 04_diversity_cagr.py and fig_category_growth (fixed with aggfunc=sum; CAGR, diversity, share CIs and growth index regenerated)","found by the Part IX writer and Agents F/G/H",executing agent (pending PI)

added taxonomy/validation/disagreements.csv +659 −0

++work_id,title,rules,rule_primary,claude-haiku-5-5,claude-sonnet-5-5,pi_review
++4400648397,Solving Third Order Linear Difference Equations in Terms of Second Order Equations,[],,['computer_algebra_placeholder'],['algorithms'],False
++4283813889,Sequential Equilibria in Signaling Games,[],,['reinforcement_learning'],['algorithms'],False
++4396572044,Multi-sensor data fusion-enabled lightweight convolutional double regularization contrast transformer for aerospace bearing small samples fault diagnosis,[],,['deep_learning_architectures'],['deep_learning_architectures'],False
++4412353443,Mapping Urban Villages in Beijing–Tianjin–Hebei Core Cities Through a Dual-Stage Fusion of Individual Building Footprints and Glimmer Imagery,[],,['computer_vision'],['computer_vision'],False
++2806158045,COMPARATIVE STUDY OF MAMDANI AND 1 ST -ORDER TAKAGI SUGENO KANG TYPE FUZZY LOGIC CONTROLLERS FOR CONTROLLING THE RIGID GANTRY CRANE SYSTEM,[],,[],['robotics'],False
++2982518672,A Stochastic Wireline Communication System,[],,['computer_architecture'],['computer_networks'],False
++2990710531,Simulation and experiment of small ball suspension height control system,[],,['robotics'],[],False
++4386952809,Relationship Between the Motives for Gaming and Subjective Well-being of University Students by Duration of Game Playing,[],,['hci'],['hci'],False
++3109176755,Proof of the Kakeya set conjecture over rings of integers modulo square-free N,[],,['computational_complexity'],[],False
++7128287921,Enhancing Textile Design Education with AI: A Case Study on Time Efficiency and Student Engagement,[],,[],['hci'],False
++3093363583,App móvil para el aprendizaje de componentes básicos de la computadora en niños de educación básica,[],,['hci'],['hci'],False
++3035934964,On nonlinear complementarity problems with applications,[],,['computational_complexity'],['algorithms'],False
++4399338544,Active and passive cooperative tracking method based on weighted least squares,[],,['computer_architecture'],[],False
++3007072699,Low Cost Portable Diagnostic Device for Automatic Classification of the Abnormal Cardiac Sound using PGC Recording,[],,['computational_biology'],['speech_audio'],False
++3007925741,Modeling Surface Roughness-Related Uncertainties of Leaky Lamb Wave Clamp-on Ultrasonic Flowmeters,[],,['computer_architecture'],[],False
++7139130213,"Modelling of cycle times for mobile racks AS/RS for functioning costs optimisation, energy costs targeted",[],,['computer_architecture'],[],False
++4409248967,Sampling and Layout of OPM Arrays for MCG Based on the Sampling Theorem,[],,['computer_architecture'],[],False
++3197510543,Changing the recent past to reduce ongoing dropout: an early learning analytics intervention for an online statistics course,[],,"['database_systems', 'information_retrieval']",[],False
++2803291160,Early Public Responses to the Zika-Virus on YouTube,[],,[],['hci'],False
++4236001208,ARMA Models,[],,['ml_theory'],[],False
++4402341510,Implementasi Metode Dempster-Shafer dalam Mendiagnosis Kelainan Neurologis Berdasarkan Perbedaan Onset Nyeri,[],,"['ai_safety', 'scientific_computing']",['ai_agents'],False
++4396949355,New Insights into the Random Telegraph Noise (RTN) in FinFETs at Cryogenic Temperature,[],,['computer_architecture'],['computer_architecture'],False
++3201351346,A Highly Efficient CMOS Rectifier for Ultra-Low-Power Ambient RF Energy Harvesting,[],,['computer_architecture'],['computer_architecture'],False
++3081174150,ВЫБОР МЕТОДОВ И СРЕДСТВ ДЛЯ ОБРАБОТКИ БОЛЬШИХ ДАННЫХ В СФЕРЕ ЭЛЕКТРОННОГО ОБУЧЕНИЯ,[],,[],['database_systems'],False
++4389301359,Evaluation and modeling of students’ persistence and wheel-spinning propensities in formative assessments,[],,['ml_theory'],[],True
++4249995757,Research on Multi-objective Optimization Scheduling of Microgrid Based on Particle Swarm Optimization Algorithm,[],,[],['algorithms'],True
++4285808510,A MOBILE TREASURE HUNT GAME FOR LANGUAGE LEARNING,[],,[],['hci'],False
++4377195986,Asynchronous event-triggered guaranteed cost control of two-dimensional Markov jump Roesser systems,[],,['ml_theory'],[],False
++4312554546,A Cross Regulation Reduced Multi-Output and Multi-VCR Piezoelectric Energy Harvesting System Using Shared Capacitors,[],,['computer_architecture'],['computer_architecture'],False
++2994665920,Towards Improved Field Application of Using Distributed Temperature Sensing for Soil Moisture Estimation: A Laboratory Experiment,[],,['computer_architecture'],[],False
++7201179105,A Model of Information Influences on the Base of Rectangular Stochastic Matrices in Chains of Reasoning with Possible Contradictions,[],,[],['ml_theory'],False
++7148607934,ATTRACTIVE - Autonomous train transponder alerting cross track intercepting vehicles,[],,"['computer_networks', 'robotics']",['robotics'],False
++4387754658,ON THE STABILITY OF SOLUTIONS FOR THE SEMI-AFFINE VARIATIONAL INEQUALITIES IN HILBERT SPACES,[],,['ml_theory'],[],False
++2943056810,Analysis of Coexistence Phenomena in Piecewise-Constant Oscillators Coupled by Hysteresis Elements,[],,['deep_learning_architectures'],[],False
++3036170140,Perancangan Sistem Informasi E-commerce pada PT Jati Jaya Nirachi Tangerang,[],,['database_systems'],['software_engineering'],False
++3159931479,Understanding employees' response to work-related after-hours use of instant messaging apps: a stress and coping perspective,[],,[],['hci'],False
++7166124027,"Empirical Estimation of Urban Photovoltaic Potential: Integrating Daily Weather Variability and Directional Scaling in Salzburg, Austria",[],,['sustainable_computing'],[],False
++2889050697,Multi-Energy Scheduling Using a Hybrid Systems Approach,[],,['ai_agents'],[],False
++4407389586,Robust and precision small-angle measurement through the speckle pattern imaging technique,[],,['computer_vision'],[],False
++4378804495,Personality recognition based on ECG in response to VR emotional stimuli,[],,['hci'],['hci'],False
++2899278391,A Machine-Learning Approach for Detection and Quantification of QRS Fragmentation,[],,['computational_biology'],['computational_biology'],False
++3217539441,Yes! I Want My Non-Cognitive Skills to Be Improved: Perceptions on an ICT-Enabled Learning Journey,[],,[],['hci'],False
++7172412385,Conjugation invariants determine the metacommutation permutation only up to relabelling,[],,['computational_complexity'],[],False
++4413050241,Development of AI Models for Early Prediction of Preterm Birth,[],,['deep_learning_architectures'],['computational_biology'],False
++2899602491,How can technology enhance elderly adherence to self-managed treatment plan?,[],,['hci'],['hci'],False
++4383695424,Joint Discrete Universality in the Selberg–Steuding Class,[],,['ml_theory'],[],False
++4405194088,Notes on variation and lexical diachrony in the Parish Memories-Alentejo collection (1758),[],,['nlp'],['nlp'],False
++4395448276,A review of the application of artificial intelligence in South African Higher Education,[],,['nlp'],[],False
++2990253946,Evacuation Supporting System Based on IoT Components †,[],,['edge_computing'],"['algorithms', 'edge_computing']",False
++4405738901,Chimera states in systems of superdiffusively coupled neurons,[],,['scientific_computing'],[],False
++4210484898,Research on Variable Parameter Drilling Method of Ti-CFRP-Ti Laminated Stacks Based on Real-Time Sensing of Drilling Axial Force,[],,['computer_architecture'],[],False
++4386047869,"Infrared hyperspectral analysis for non-invasive, inline fat content determination in bakery products",[],,['computer_vision'],['computer_vision'],False
++3168268628,Supporting Collaboration with Forgetting Curves,[],,['information_retrieval'],['hci'],False
++2947916735,Multicriteria Model of Students’ Knowledge Diagnostics Based on the Doubt Measuring Level Method for E&M Learning,[],,['hci'],['hci'],False
++7148627082,Méthode d'optimisation mixte bio-inspirée : application à l'imagerie multi-spectrale et à la mesure d'audience,[],,"['computer_vision', 'deep_learning_architectures']","['algorithms', 'computer_vision']",False
++4352989561,Implementation of a Social Network Information Dissemination Model Incorporating Negative Relationships,[],,['database_systems'],['algorithms'],False
++4406799739,A Study on Mixed Variational-like Inequality Under Generalized Quasimonotonicity in Banach Spaces,[],,[],['algorithms'],False
++3006452720,A lightweight ensemble spatiotemporal interpolation model for geospatial data,[],,['database_systems'],['scientific_computing'],False
++4388634849,‘I pretend to be an ideal woman just to keep their mouths shut’: Bangladeshi women’s contestation of abuse through social media platforms,[],,[],['hci'],False
++4411680001,Innovative Research on the Role of Artificial Intelligence in Cultivating Imagination in Children's Creative Art Activities,[],,['hci'],['hci'],False
++3206918036,Copernicus Global Land Service Quality Assessment – Better Good Than Sorry!,[],,['computer_vision'],[],False
++4412965306,Bidirectional Long Short-Term Memory Approach for Infrastructure Health Monitoring Using On-Board Vibration Response,[],,['deep_learning_architectures'],['deep_learning_architectures'],False
++4415671687,Geostatistically Enhanced Learning for Supervised Classification of Wall-Rock Alteration Using Assay Grades of Trace Elements and Sulfides,[],,['ml_theory'],['ml_theory'],True
++3160907559,Internet of Things architecture challenges: A systematic review,"['computer_networks', 'edge_computing']",computer_networks,['edge_computing'],"['computer_networks', 'software_engineering']",False
++4210382415,Communication-Efficient Coded Distributed Multi - Task Learning,"['computer_networks', 'cybersecurity', 'ml_theory', 'privacy_preserving_computation']",computer_networks,"['distributed_computing', 'ml_theory']","['distributed_computing', 'federated_learning']",False
++4389101774,Wireless Bilateral Breath Monitoring System Using Temperature-Based Wearable Technology in Motion,['computer_networks'],computer_networks,"['computer_networks', 'hci']",['hci'],False
++2775587204,The Role of Social Networking Sites as Information Sources in the Educational - Learning Process from the Perspective of a Sample of Jordan University Students,['computer_networks'],computer_networks,[],[],False
++3035035809,Representing Joint Hierarchies with Box Embeddings,"['deep_learning_architectures', 'llm', 'nlp']",nlp,['ml_theory'],"['deep_learning_architectures', 'nlp']",False
++4285254984,An impact of ontology-based service-oriented ecosystems on digital transformation of railway transport and engineering education,['nlp'],nlp,['hci'],[],False
++4361006333,"Root cause prediction for failures in semiconductor industry, a genetic algorithm–machine learning approach","['compilers', 'ml_theory', 'nlp']",nlp,"['computer_architecture', 'ml_theory']",['nlp'],False
++4399923297,IMPLEMENTASI CHAT BOT BERBASIS TEORI KEPRIBADIAN HIPPOCRATES YANG TERINTEGRASI WEB,['nlp'],nlp,['hci'],"['hci', 'nlp']",False
++7169949244,Análisis de la Sostenibilidad en la Moda. Análisis de Sentimientos en los Comentarios de YouTube,['nlp'],nlp,['nlp'],[],False
++7204475324,"Students’ use of artificial intelligence in higher education, and their reflections on that use: an interview study",['nlp'],nlp,[],['hci'],False
++2948072807,Electronically Tunable Reflection Type Phase Shifters,['hci'],hci,['computer_architecture'],[],False
++7171785185,Insights from Eye-Tracking Technology in Infants with or at Risk of a Motor Disability: A Systematic Review,"['computer_vision', 'database_systems', 'hci']",hci,[],[],False
++4403160367,Evaluating Environments Using Exploratory Agents,['hci'],hci,"['computer_graphics', 'hci', 'reinforcement_learning']",['ai_agents'],False
++7135288805,L'impact de réseau social Facebook sur l’écriture des jeunes algériens,['hci'],hci,[],[],False
++7220333334,Criterios para el diseño de un sitio de Ecommerce (B2C),['hci'],hci,['hci'],[],False
++4409620177,Students' Perception on The Use of Pear Deck on English Learning,['hci'],hci,[],[],False
++2980067317,Improving The Performance of Motor Imagery Based Brain-Computer Interface Using Phase Space Reconstruction,['hci'],hci,"['ai_safety', 'speech_audio']",[],False
++3166564139,WEB-BASED JOB TRAINING INFORMATION SYSTEM FOR ELECTRICAL ENGINEERING STUDY PROGRAM STUDENTS,"['database_systems', 'hci']",hci,['database_systems'],[],False
++2951084085,Scaling-Out Longitudinal Clinical Analytics with Dataflow Processing,"['cloud_computing', 'computer_vision', 'database_systems', 'hci']",hci,"['distributed_computing', 'robotics']","['computational_biology', 'distributed_computing']",False
++7198871956,Tengger Communities Creativity and Language Preservation in Text-Speak on the WhatsApp Social Media Platform,['hci'],hci,[],[],False
++2911964379,Quality Control in Crowdsourcing Using Sequential Zero-Determinant Strategies,"['distributed_computing', 'hci', 'information_retrieval']",hci,"['information_retrieval', 'reinforcement_learning']",['algorithms'],False
++4285312412,Nociones de shock y atención en videos de Tik Tok,"['applied_cryptography', 'hci']",hci,[],[],False
++4413901110,Aplicación de la inteligencia artificial para mejorar la comprensión lectora y escritura en estudiantes secundarios,['hci'],hci,[],[],False
++7123361757,Emotional Experiences in Virtual Reality: Effects of Immersive Environments on Emotional Responses,['hci'],hci,[],['hci'],False
++7161768941,"English address terms in Australian, British and North American English on Twitter/X",['hci'],hci,[],[],False
++4413188996,BRECHA DIGITAL EN LA ENSEÑANZA: LA IMPERIOSA NECESIDAD DE FORMACIÓN TECNOLÓGICA DOCENTE PARA LA EDUCACIÓN DEL SIGLO XXI,['hci'],hci,[],[],False
++7216627434,PELATIHAN PEMBUATAN DAN PENGGUNAAN MATEMATIKA KREATIF AUGMENTED REALITY SEBAGAI MEDIA PEMBELAJARAN GEOMETRI BAGI GURU DAN MAHASISWA PGSD,['hci'],hci,[],[],False
++4400230898,An Intracortical Wireless Bidirectional Brain-Computer Interface with High Data Density,['hci'],hci,['computer_architecture'],['computer_architecture'],False
++4404037745,SPIR with Colluding and Non-Replicated Servers from a Noisy Channel,"['applied_cryptography', 'cybersecurity', 'information_retrieval', 'privacy_preserving_computation']",privacy_preserving_computation,"['applied_cryptography', 'computational_complexity']",['privacy_preserving_computation'],False
++7214618239,Persona-Based Adversarial Fraud Transaction Generation Framework for Dynamic Evaluation of AI-Based Fraud Detection Systems,['privacy_preserving_computation'],privacy_preserving_computation,"['ai_safety', 'cybersecurity', 'generative_ai', 'nlp']","['cybersecurity', 'llm']",False
++7165521315,Optimal mechanisms for selling market segmentation,['privacy_preserving_computation'],privacy_preserving_computation,['ml_theory'],['algorithms'],False
++7125647931,ЗМІЦНЕННЯ ІНФОРМАЦІЙНОЇ БЕЗПЕКИ ТА КОНФІДЕНЦІЙНОСТІ ПЕРСОНАЛЬНИХ ДАНИХ З ВИКОРИСТАННЯМ ШІ,['privacy_preserving_computation'],privacy_preserving_computation,[],[],False
++3093013316,"COVID-19, Public Management and Type 1 Errors","['cybersecurity', 'privacy_preserving_computation']",privacy_preserving_computation,[],['privacy_preserving_computation'],False
++3012891874,Statistical Analysis of Identity Risk of Exposure and Cost Using the Ecosystem of Identity Attributes,"['cybersecurity', 'privacy_preserving_computation']",privacy_preserving_computation,"['database_systems', 'privacy_preserving_computation']",['cybersecurity'],False
++4415149235,Target and Non-target Category Classification from GPS and Check-in Data,"['cybersecurity', 'privacy_preserving_computation']",privacy_preserving_computation,"['information_retrieval', 'ml_theory']",['database_systems'],False
++7155107746,Efficient Synthetic Data Selection via Pontryagin’s Maximum Principle,"['computational_complexity', 'privacy_preserving_computation']",privacy_preserving_computation,"['federated_learning', 'generative_ai']","['generative_ai', 'privacy_preserving_computation']",False
++4382045981,“It may be a pain in the backside but...” Insights into the resilience of business after GDPR,['privacy_preserving_computation'],privacy_preserving_computation,[],['privacy_preserving_computation'],False
++4388951965,The study of PGP web of trust based on social network analysis,"['applied_cryptography', 'privacy_preserving_computation']",privacy_preserving_computation,"['cybersecurity', 'information_retrieval']",['cybersecurity'],False
++7165783687,"When AI Conversations Become Advertising Data: Algorithmic Trust, Privacy Calculus, and Purchase Intention in GenAI-Personalized Social Commerce","['algorithms', 'generative_ai', 'privacy_preserving_computation']",privacy_preserving_computation,['hci'],['hci'],False
++2786217712,Freedom of Encryption,"['applied_cryptography', 'privacy_preserving_computation']",privacy_preserving_computation,['cybersecurity'],['privacy_preserving_computation'],True
++4386192152,Development of Expert System Application to Detect Chicken Disease using the Forward Chaining Method,['privacy_preserving_computation'],privacy_preserving_computation,['ai_safety'],[],False
++3011050248,Sensor Signatures: AI Assisted Co-Location Detection of IoT Devices,"['computer_networks', 'computer_vision', 'edge_computing']",edge_computing,"['edge_computing', 'ml_theory']",['computer_networks'],False
++3090747742,Modeling Product Manufacturing Reliability with Quality Variations Centered on the Multilayered Coupling Operational Characteristics of Intelligent Manufacturing Systems,['edge_computing'],edge_computing,[],[],False
++4214663994,Notice of Violation of IEEE Publication Principles: Affirmative Fusion Process for Improving Wearable Sensor Data Availability in Artificial Intelligence of Medical Things,"['computer_networks', 'edge_computing']",edge_computing,['computational_biology'],['computational_biology'],False
++4389319530,Preface,"['computer_networks', 'edge_computing', 'scientific_computing']",edge_computing,[],[],False
++3007216950,اینترنت اشیاء و تعامل با دنیای فیزیکی,['edge_computing'],edge_computing,['computer_networks'],[],False
++4385153790,A Novel Accurate Insecticidal Counting Method Based on Solar Insecticidal Lamp Using Machine Learning,['edge_computing'],edge_computing,"['computer_architecture', 'edge_computing']",[],True
++4312555758,IoT Based Pre-Operative Prehabilitation Program Monitoring Model: Implementation and Preliminary Evaluation,"['computer_vision', 'edge_computing', 'hci']",edge_computing,"['computational_biology', 'edge_computing']",['hci'],False
++3199685485,A Fast Wake-Up and High Accuracy Sensor Interface by Synchronous Sampling with Power-Efficient Switching Regulator,['edge_computing'],edge_computing,['computer_architecture'],['computer_architecture'],False
++4401883192,Machine Learning-Driven Smart Discovery in IoT-Enabled Environments,"['computer_networks', 'cybersecurity', 'edge_computing', 'privacy_preserving_computation']",edge_computing,['ml_theory'],[],False
++3184417063,CALM: Contactless Accurate Load Monitoring via Modality Distillation,['edge_computing'],edge_computing,"['ml_theory', 'speech_audio']",['deep_learning_architectures'],False
++7156021111,Analysis of the data processing method and system design of the IoT hardware gateway using a microcontroller,"['computer_architecture', 'computer_networks', 'edge_computing', 'nlp']",edge_computing,"['edge_computing', 'ml_theory']",[],False
++4411194662,Security-Based Distributionally Robust Optimization Self-Triggered SMPC for Constrained Cyber–Physical Systems Subject to Unknown Disturbances and Denial-of-Service Attacks,"['cybersecurity', 'edge_computing', 'scientific_computing']",edge_computing,"['ai_safety', 'ml_theory']",[],False
++4281719112,NanoElectroOptical Nose (NEONOSE) for the detection of Climate Change gases,"['computer_architecture', 'edge_computing']",edge_computing,['computer_architecture'],[],False
++4410459414,A Literature Review Framework and Open Research Challenges for Predictive Maintenance in industry 4.0,['edge_computing'],edge_computing,"['database_systems', 'ml_theory']",[],False
++4200300237,Millions of records leaked as T-Mobile is breached for the sixth time in four years,['cybersecurity'],cybersecurity,[],['cybersecurity'],False
++3089767340,That password doesn't sound right,['cybersecurity'],cybersecurity,['hci'],"['cybersecurity', 'hci']",False
++2995203890,New Image Watermarking Algorithm Based on DWT and Pixel Movement Function PMF,"['computer_vision', 'cybersecurity', 'scientific_computing']",cybersecurity,['computer_vision'],['cybersecurity'],False
++4393105534,Media Pembelajaran Seks Untuk Remaja,['cybersecurity'],cybersecurity,[],[],False
++7118011917,Functions,"['programming_languages', 'software_engineering']",software_engineering,['software_engineering'],['programming_languages'],False
++4253254765,Working with Tables,['software_engineering'],software_engineering,[],[],True
++4386014351,Framework for usage potential for Digital Procurement in the Construction Industry,['software_engineering'],software_engineering,['information_retrieval'],[],False
++4280590664,Business Transformation Project’s Holistic Agile Management (BTPHAM),"['programming_languages', 'software_engineering']",software_engineering,[],['software_engineering'],False
++4312839447,Building,['software_engineering'],software_engineering,[],['software_engineering'],False
++4206446312,Decentralized Holonic Control System Model for Line-less Mobile Assembly Systems,['software_engineering'],software_engineering,['ai_agents'],"['ai_agents', 'robotics']",False
++7203691417,Effectiveness of lazy algorithm-based software effort estimation,['software_engineering'],software_engineering,['ml_theory'],['software_engineering'],False
++7162132995,ConTact: Contact-First Antibody CDR Design via Explicit Interface Reasoning,"['distributed_computing', 'software_engineering']",software_engineering,"['computational_biology', 'deep_learning_architectures']","['computational_biology', 'deep_learning_architectures']",False
++2908635758,Paradoxical Simulations to Enhance Education in Mathematics,"['scientific_computing', 'software_engineering']",software_engineering,['nlp'],[],False
++7113083064,"Generative AI in Programming Education : An Interpretive, Sociotechnical Analysis of Student and Teacher Perspectives","['generative_ai', 'llm', 'software_engineering']",software_engineering,['hci'],['hci'],True
++7163811937,Befitting Bootstrap Analysis: The Effect of Imbalance,"['operating_systems', 'scientific_computing', 'software_engineering']",software_engineering,['ml_theory'],['ml_theory'],False
++2931497058,Aplikasi Pemandu Wisata Kabupaten Kepulauan Selayar Menggunakan Google Maps API Pada Platform Android,"['database_systems', 'software_engineering']",software_engineering,['hci'],['software_engineering'],True
++4382935921,On the design of multiplex control to reject disturbances in nonlinear network systems affected by heterogeneous delays,['robotics'],robotics,['ml_theory'],[],False
++3034448848,A Fault Detection and Isolation Design for a Dual Pitot Tube Air Data System,['robotics'],robotics,['computer_architecture'],['robotics'],False
++4394880330,Estimation of All-Day Aerosol Optical Depth in the Beijing–Tianjin–Hebei Region Using Ground Air Quality Data,['robotics'],robotics,['scientific_computing'],['computer_vision'],False
++4307184398,Preservice teachers’ opinion about developing computational thinking in preschool,['robotics'],robotics,['hci'],[],True
++4313227226,Multi-UAV coordinated track planning method based on MSISOS algorithm,['robotics'],robotics,['ai_agents'],"['algorithms', 'robotics']",False
++4382059457,Some agents are more similar than others: customer orientation of frontline robots and employees,"['hci', 'robotics']",robotics,['hci'],[],False
++4410294818,Regularization for Deep Imbalanced Regression Based on Quantitative Relationship,"['computer_vision', 'ml_theory']",ml_theory,"['deep_learning_architectures', 'ml_theory']",['deep_learning_architectures'],False
++7215048402,Comparative evaluation of resampling techniques for improving the performance of classification algorithms on a breast cancer dataset,['ml_theory'],ml_theory,[],[],False
++7219508654,Classification of Medical Students' Skills Using the Random Forest Method Based on Practicum and Theory Scores,"['computer_architecture', 'ml_theory']",ml_theory,[],[],False
++3155002205,Accurate and fast matrix factorization for low-rank learning,"['computer_vision', 'deep_learning_architectures', 'information_retrieval', 'ml_theory', 'scientific_computing']",ml_theory,['algorithms'],['scientific_computing'],False
++4327500503,Anomalous/Relevant Event Detection (A/RED): Active Machine Learning for Finding Rare Events,"['database_systems', 'ml_theory']",ml_theory,['ml_theory'],[],False
++7171392614,A Bayesian-Optimized Deep Scaled Stacking Framework for Data-Driven Intelligent Control in Continuous Oil Refining Process,"['applied_cryptography', 'ml_theory', 'operating_systems']",ml_theory,['ml_theory'],[],False
++4402510141,Approximations of system W for inference from strongly and weakly consistent belief bases,['ml_theory'],ml_theory,['ml_theory'],[],False
++4388726461,Should I Sample it or Not? Improving Quality Assurance Efficiency Through Smart Active Sampling,"['ml_theory', 'software_engineering']",ml_theory,['ml_theory'],[],False
++7170116302,Artificial neural network modelling of student’s academic performance based on online learning activities pattern,"['deep_learning_architectures', 'ml_theory']",ml_theory,['ml_theory'],[],False
++4385985189,A Spatial Data-Driven Approach for Mineral Prospectivity Mapping,['ml_theory'],ml_theory,['ml_theory'],[],False
++7151296083,Toward unified anomaly detection with an application to fraud detection,['ml_theory'],ml_theory,['computer_vision'],['computer_vision'],False
++4408365506,Construction of Western Music Theory Teaching Model Based on Machine Learning,"['ml_theory', 'speech_audio']",ml_theory,[],[],False
++3012094357,Anomaly Detection Method for Operation and Maintenance Data Based on One-Class Learning,['ml_theory'],ml_theory,"['database_systems', 'ml_theory']",[],True
++4392903168,Unsupervised Anomaly Detection for Multivariate Time Series Using Diffusion Model,"['ml_theory', 'software_engineering']",ml_theory,"['deep_learning_architectures', 'generative_ai']","['deep_learning_architectures', 'generative_ai']",False
++4234642263,Design and Development of Financial Fraud Detection using Machine Learning,"['ml_theory', 'privacy_preserving_computation']",ml_theory,['ml_theory'],[],False
++7213442406,Evaluating EMOO as a Metric-Balanced Ensemble Framework for Multi-Objective Optimization in Imbalanced Medical Data,"['algorithms', 'ml_theory', 'software_engineering']",ml_theory,"['computational_biology', 'ml_theory']",['computational_biology'],False
++3027701685,Detecting Anomalies in Application Performance Management System with Machine Learning Algorihms,"['computer_networks', 'cybersecurity', 'ml_theory']",ml_theory,"['cybersecurity', 'database_systems']",[],True
++4396611119,Machine Learning and Weather Model Combination for PV Production Forecasting,"['ml_theory', 'scientific_computing']",ml_theory,['ml_theory'],[],False
++3005800578,My Continuing Journey from Logic to Computational Argumentation,"['ai_agents', 'formal_verification']",formal_verification,[],['ai_agents'],False
++4403534925,A Model of Predictive Profit Management Policy,['formal_verification'],formal_verification,[],[],False
++2904264076,First-Principles Design of Refractory High Entropy Alloy VMoNbTaW,['formal_verification'],formal_verification,['scientific_computing'],[],False
++4405491888,Modeling and Performance Assessment of Drone-based Food Delivery Logistics Using Colored Generalized Stochastic Petri Nets,"['formal_verification', 'scientific_computing']",formal_verification,['scientific_computing'],['formal_verification'],False
++2985419616,The Semantics of Propositions. Truth and Consequences,['formal_verification'],formal_verification,[],[],False
++2998652541,Fusing Multi-Sensor Measurements to Improve Heading Estimation using Kalman Gain for Indoors Smartphone Positioning Solutions,['formal_verification'],formal_verification,"['computer_vision', 'hci']",['robotics'],False
++4292072122,Max-plus polyhedra-based state characterization for uMPL systems,['formal_verification'],formal_verification,['ml_theory'],['algorithms'],True
++2919546362,Method to Improve Accuracy of Indoor PDR Trajectories Using a Large Amount of Trajectories,['formal_verification'],formal_verification,['hci'],['hci'],False
++7202355676,Использование сетей Петри для синхронизации процессов пополнения парка изотермических транспортных средств в период повышенного сезона,['formal_verification'],formal_verification,['algorithms'],[],False
++2941019358,Circuit Relations for Real Stabilizers: Towards TOF+H,['formal_verification'],formal_verification,"['formal_verification', 'quantum_computing']",['quantum_computing'],False
++2888263960,Effects of Nitrogen Content on the Structure and Mechanical Properties of (Al0.5CrFeNiTi0.25)Nx High-Entropy Films by Reactive Sputtering,['formal_verification'],formal_verification,[],[],False
++2980088699,Tangent Categories from the Coalgebras of Differential Categories,"['formal_verification', 'programming_languages']",formal_verification,['formal_verification'],['programming_languages'],False
++2969344391,MOCA: An ORM model for computational accountability,"['ai_agents', 'distributed_computing', 'formal_verification']",formal_verification,"['ml_theory', 'software_engineering']",['software_engineering'],False
++4414118222,Similarity-Based Constraint Solving Modulo Commutativity: A Unified Framework,"['formal_verification', 'programming_languages']",formal_verification,['algorithms'],"['formal_verification', 'programming_languages']",False
++4391021938,Interval Observers for Hybrid Dynamical Systems with Known Jump Times,['formal_verification'],formal_verification,['ml_theory'],[],False
++4402325914,Improving the Resilience of Barrier Function-Based Safety-Critical Controllers Against Uncertainties with Application to Connected Autonomous Vehicles,['formal_verification'],formal_verification,['robotics'],['robotics'],False
++3158571805,A receding horizon event-driven control strategy for intelligent traffic management,['formal_verification'],formal_verification,"['ai_agents', 'robotics']",['formal_verification'],False
++4415180303,Quantification of Sim2Real Gap via Neural Simulation Gap Function,"['formal_verification', 'reinforcement_learning', 'scientific_computing']",formal_verification,"['ml_theory', 'robotics']","['ml_theory', 'robotics']",False
++4400077792,Delayed Decision Guidance in an Integrated Midcourse-Terminal Engagement for Conventional Interceptors,"['cybersecurity', 'formal_verification']",formal_verification,['robotics'],['robotics'],False
++3116998496,Gait data analysis based on acceleration sensor,['computer_vision'],computer_vision,['robotics'],[],False
++4410134551,The Second Renaissance,"['autonomous_scientific_discovery', 'cybersecurity', 'scientific_computing']",autonomous_scientific_discovery,[],[],False
++7168123153,Emerging Artificial Intelligence in Mathematics Education: A Systematic Literature Review,['autonomous_scientific_discovery'],autonomous_scientific_discovery,['nlp'],[],False
++3168036902,Dynamic Design of the Turbomachinery Blade with the Joint Application of the Newly Developed Constrained Zone and the PSO Algorithm,['autonomous_scientific_discovery'],autonomous_scientific_discovery,['computer_graphics'],['algorithms'],False
++4318620721,The Automated Discovery of Kinetic Rate Models -- Methodological Frameworks,"['autonomous_scientific_discovery', 'computational_biology', 'computational_complexity', 'scientific_computing']",autonomous_scientific_discovery,['scientific_computing'],['autonomous_scientific_discovery'],False
++7169848785,Preface for the Symbolic and Generative AI for Science,"['autonomous_scientific_discovery', 'generative_ai']",autonomous_scientific_discovery,[],['autonomous_scientific_discovery'],True
++3003447096,"Raw-to-repository characterization data conversion for repeatable, replicable, and reproducible measurements","['autonomous_scientific_discovery', 'computer_networks', 'distributed_computing', 'scientific_computing']",autonomous_scientific_discovery,['database_systems'],['database_systems'],False
++4410289386,Exploring the nexus of many-body theories through neural network techniques: the tangent model,['autonomous_scientific_discovery'],autonomous_scientific_discovery,['scientific_computing'],['scientific_computing'],False
++4405426097,Quantifying Behavioral Dissimilarity Between Mathematical Expressions,"['autonomous_scientific_discovery', 'software_engineering']",autonomous_scientific_discovery,['ml_theory'],"['autonomous_scientific_discovery', 'ml_theory']",True
++7214605398,The B-AIMT: development and initial validation of a domain-specific instrument for assessing mathematics teachers’ beliefs about artificial intelligence,['autonomous_scientific_discovery'],autonomous_scientific_discovery,['hci'],[],True
++4399747476,FUTURISTIC TRENDS IN ARTIFICIAL INTELLIGENCE,"['ai_safety', 'autonomous_scientific_discovery', 'computer_vision']",autonomous_scientific_discovery,[],[],False
++7113026192,Exploring Fraction Comprehension and Interest in Elementary Education Through AI-Powered Personalized Learning,['autonomous_scientific_discovery'],autonomous_scientific_discovery,['hci'],[],False
++7117871859,Exploring the Potential of Artificial Intelligence Tools in Teaching and Learning: Mathematics Education Considerations,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],True
++7139945210,$R$-equivalence on Cubic Surfaces I: Existing Cases with Non-Trivial Universal Equivalence,"['autonomous_scientific_discovery', 'generative_ai', 'hci']",autonomous_scientific_discovery,[],['autonomous_scientific_discovery'],False
++4411799567,A scientific future shared with AI,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],['autonomous_scientific_discovery'],False
++4414528656,SELF-DIRECTED LEARNING WITH YOUTUBE: ENHANCING LISTENING SKILLS ON STUDENTS OF ENGLISH EDUCATION STUDY PROGRAM,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],False
++7154965180,Emergent symmetry in mushroom-based foods,"['autonomous_scientific_discovery', 'scientific_computing']",autonomous_scientific_discovery,[],[],False
++4403160419,Complete and Efficient Covariants for 3D Point Configurations with Application to Learning Molecular Quantum Properties,['autonomous_scientific_discovery'],autonomous_scientific_discovery,"['computational_biology', 'deep_learning_architectures']","['deep_learning_architectures', 'ml_theory']",False
++2896073529,2018 ACM SIGAI student essay contest on artificial intelligence technologies,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],False
++7207633319,Pemanfaatan Artificial Intelligence (AI) dalam Proses Pembelajaran IPA di Kelas Rendah pada Tema Panca Indra untuk Memotivasi Siswa UPT SDN 01 Mekar Laras,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],True
++7127595070,Student Workshop on Scientific Writing – A Practical Approach to Workshop Design,['autonomous_scientific_discovery'],autonomous_scientific_discovery,['nlp'],[],False
++3131458733,Predicting the determinants of online learning adoption during the COVID-19 outbreak: a two-staged hybrid SEM-neural network approach,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],False
++7121660003,Research on Business Design Decisions Driven by Machine Learning,"['autonomous_scientific_discovery', 'ml_theory']",autonomous_scientific_discovery,"['information_retrieval', 'ml_theory']",[],False
++7204883563,DS-Lighting: Making Agent Harnesses Explicit for Data-Science Automation,"['autonomous_scientific_discovery', 'llm', 'software_engineering']",autonomous_scientific_discovery,"['ai_agents', 'llm']",['ai_agents'],False
++7163871521,Perspectives on South Africa’s open science policy: artificial intelligence and data ownership as unresolved questions,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],['ai_safety'],False
++4200490276,Automated Machine Learning Pipeline System Based on Beam Search for Electric Power Industry,"['autonomous_scientific_discovery', 'blockchain', 'ml_theory', 'nlp']",autonomous_scientific_discovery,"['database_systems', 'ml_theory']",[],False
++7165673775,Implementation of Intelligent Systems and Digital Technologies in Teaching Engineering Disciplines,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],False
++7217524975,การพฒนาบทเรยนคอมพวเตอรชวยสอนตามรปแบบการเรยนรดวยตนเอง วชา โปรแกรมนำเสนอ เรองการใชงานโปรแกรม Microsoft PowerPoint สำหรบนกเรยนชนมธยมศกษา ปท 6 โรงเรยนคระบรชยพฒนาพทยาคม,['autonomous_scientific_discovery'],autonomous_scientific_discovery,[],[],False
++4224247694,Enabling mechano-intelligence in adaptive structures utilizing physical computing,['deep_learning_architectures'],deep_learning_architectures,"['deep_learning_architectures', 'robotics']",['ml_theory'],False
++4410956047,Power Forecasting of Distributed Photovoltaic Systems Based on EEMD-FCMs,['deep_learning_architectures'],deep_learning_architectures,['scientific_computing'],['deep_learning_architectures'],False
++3011306938,Link Prediction using Graph Neural Networks for Master Data Management,"['database_systems', 'deep_learning_architectures', 'privacy_preserving_computation']",deep_learning_architectures,"['database_systems', 'information_retrieval']","['deep_learning_architectures', 'privacy_preserving_computation']",False
++2978220519,The reservoir’s perspective on generalized synchronization,['deep_learning_architectures'],deep_learning_architectures,"['deep_learning_architectures', 'ml_theory']",['ml_theory'],False
++4312870744,Fast Multipliers for Hardware Implementation of Artificial Neural Networks,"['computer_architecture', 'deep_learning_architectures', 'scientific_computing']",deep_learning_architectures,"['ai_accelerators', 'computer_architecture']",['ai_accelerators'],False
++4403331393,Optimal Power Flow Using PSO Algorithms Based on Artificial Neural Networks,['deep_learning_architectures'],deep_learning_architectures,['scientific_computing'],[],False
++7126095475,COMPARATIVE ANALYSIS OF MACHINE LEARNING AND DEEP LEARNING APPROACHES FOR PREDICTING STUDENT DROPOUT IN HIGHER EDUCATION: A CASE STUDY OF THE VIRTUAL UNIVERSITY OF IVORY COAST,"['deep_learning_architectures', 'nlp']",deep_learning_architectures,"['database_systems', 'ml_theory']",[],False
++4400948005,A Comparative Study of Multi-Factor Quantitative Investment Models Based on Investor Sentiment,"['blockchain', 'deep_learning_architectures', 'nlp']",deep_learning_architectures,"['nlp', 'reinforcement_learning']",[],True
++4416583002,BiomSHARP: Biomass Super-Resolution for High Accuracy Prediction,"['deep_learning_architectures', 'scientific_computing', 'software_engineering']",deep_learning_architectures,"['computer_vision', 'scientific_computing']","['computer_vision', 'deep_learning_architectures']",False
++4402273825,Novel stability criterion for DNNs via improved asymmetric LKF,['deep_learning_architectures'],deep_learning_architectures,"['deep_learning_architectures', 'ml_theory']",['ml_theory'],False
++2797984158,Semi-supervised Multi-kernel Extreme Learning Machine,['deep_learning_architectures'],deep_learning_architectures,"['deep_learning_architectures', 'ml_theory']",['ml_theory'],False
++7141530726,An Electrochemical Single Particle Model-Based State of Charge Estimation Using Orthogonal Collocation Levenberg Marquardt Backpropagation Hybrid Neural Network,"['computer_architecture', 'cybersecurity', 'deep_learning_architectures']",deep_learning_architectures,"['computational_biology', 'deep_learning_architectures']",['scientific_computing'],False
++4391810135,A Model of Multi-Finger Coordination in Keystroke Movement,"['algorithms', 'computer_graphics', 'deep_learning_architectures', 'robotics']",deep_learning_architectures,"['computational_biology', 'hci']",[],False
++4393372557,Classifying Functional Brain Graphs Using Graph Hypervector Representation,"['deep_learning_architectures', 'edge_computing']",deep_learning_architectures,"['computational_biology', 'deep_learning_architectures']","['computational_biology', 'edge_computing']",False
++4414859951,Demonstration of reservoir computing in a double-drum resonator through sideband pumping approach,['deep_learning_architectures'],deep_learning_architectures,"['ai_accelerators', 'computer_architecture']",[],False
++7220870911,Peer-to-peer va ierarxik tarmoqlar,['deep_learning_architectures'],deep_learning_architectures,[],[],True
++7109160285,An Energy-Based Mechanism for Compositional Behavior,"['deep_learning_architectures', 'reinforcement_learning']",deep_learning_architectures,"['ml_theory', 'reinforcement_learning']","['ai_agents', 'ml_theory']",False
++4410775015,Analysis of Hypertuning Based CNN Breast Cancer Predictor for Different Random States,['deep_learning_architectures'],deep_learning_architectures,"['computational_biology', 'computer_vision']",['deep_learning_architectures'],False
++3047152076,The World as a Neural Network,['deep_learning_architectures'],deep_learning_architectures,"['deep_learning_architectures', 'ml_theory']",['ml_theory'],False
++4283646524,Deep-Learning for the Diagnosis of Esophageal Cancers and Precursor Lesions in Endoscopic Images: A Model Establishment and Nationwide Multicenter Performance Verification Study,['deep_learning_architectures'],deep_learning_architectures,"['computational_biology', 'computer_vision']",['computer_vision'],False
++3183119475,Co-evolution of Opinion and Social Tie Dynamics Towards Structural Balance,['algorithms'],algorithms,['ml_theory'],['algorithms'],False
++2895538521,Conjugate Laplacian matrices of a graph,['algorithms'],algorithms,['algorithms'],[],False
++4401726474,Block-Sparse Tensor Recovery,"['algorithms', 'computational_complexity']",algorithms,['ml_theory'],['ml_theory'],False
++4400392879,La grande épopée des algorithmes,['algorithms'],algorithms,[],[],False
++3200096258,Degrees of Kalman varieties of tensors,['algorithms'],algorithms,['computational_complexity'],['scientific_computing'],False
++4392544619,Vector-Matrix Reversal Operation,['algorithms'],algorithms,[],[],False
++2981962366,Analyzing students' recontextualization strategies for algorithmic concepts,['algorithms'],algorithms,['hci'],[],False
++4396597290,Adaptive Cooperative Game Penetration Guidance for Multiple Hypersonic Vehicles With Performance Constraints,['algorithms'],algorithms,"['reinforcement_learning', 'robotics']",['reinforcement_learning'],False
++4409058110,Remaining Useful Life Prediction Methodologies With Health Indicator Dependence for Rotating Machinery: A Comprehensive Review,['algorithms'],algorithms,[],[],False
++3044771921,Omnipresent yet elusive: Teachers’ views on contexts for teaching algorithms in secondary education,['algorithms'],algorithms,[],[],True
++4410943262,Optimization and Comparative Analysis of Industrial Polymerization Reactors Using Physics-Inspired Metaheuristics,['algorithms'],algorithms,['scientific_computing'],['algorithms'],False
++2758153130,Pathwise Dynamic Programming,['algorithms'],algorithms,['scientific_computing'],['algorithms'],False
++3123947211,Jacobian Consistency of a Smoothing Function for the Weighted Second-Order Cone Complementarity Problem,"['algorithms', 'computer_architecture', 'cybersecurity', 'scientific_computing']",algorithms,['algorithms'],['scientific_computing'],False
++4415250135,Optimizing Sparse Antenna Arrays for Localization and Sensing using Vector Spherical Wave Functions,"['algorithms', 'edge_computing']",algorithms,['computer_architecture'],[],False
++7149228871,Metaphors in LLMs,"['llm', 'nlp']",llm,[],"['llm', 'nlp']",False
++4416889670,Future Knowledge Era (Next AI Generation Era),"['computational_biology', 'deep_learning_architectures', 'generative_ai', 'hci', 'llm', 'quantum_computing']",llm,[],[],False
++7214900323,ОСОБЛИВОСТІ ВИКОРИСТАННЯ ВЕЛИКИХ МОВНИХ МОДЕЛЕЙ СТУДЕНТАМИ НЕФІЛОЛОГІЧНИХ СПЕЦІАЛЬНОСТЕЙ У ПРОЦЕСІ ВИВЧЕННЯ ІНОЗЕМНОЇ МОВИ,"['generative_ai', 'llm', 'speech_audio']",llm,[],[],False
++7171730128,Human Preference aligned Tabular Similarity,"['database_systems', 'foundation_models', 'information_retrieval', 'llm']",llm,['ml_theory'],"['database_systems', 'information_retrieval']",False
++4387819913,Investigating Effective Ways to Use Artificial Intelligence in Teacher Education,"['llm', 'nlp']",llm,['hci'],['hci'],False
++4414992583,Artificial Intelligence as a Pedagogical Scaffold in K–12 Language and Literacy Education,"['federated_learning', 'llm', 'nlp']",llm,[],[],False
++4387935330,LEGAL ASPECTS OF INTRODUCING ARTIFICIAL INTELLIGENCE IN INCLUSIVE EDUCATION,['llm'],llm,[],[],False
++7199029551,The ambiguity of searching for linguistic representations in large language models,"['llm', 'nlp']",llm,[],['llm'],True
++4312852711,A data-driven medical knowledge discovery framework to predict the length of ICU stay for patients undergoing craniotomy based on electronic medical records,['computational_biology'],computational_biology,"['deep_learning_architectures', 'ml_theory']",['computational_biology'],False
++3127450329,Association between evidence-based training and clinician proficiency in electronic health record use,['computational_biology'],computational_biology,['hci'],['hci'],False
++4391930238,A Comprehensive Review of Machine Learning Techniques for Lung Cancer Detection,['computational_biology'],computational_biology,['computer_vision'],['computer_vision'],False
++4411492298,Shinyscreen: mass spectrometry data inspection and quality checking utility,"['cloud_computing', 'computational_biology', 'computer_networks', 'database_systems', 'hci', 'operating_systems']",computational_biology,['scientific_computing'],['computational_biology'],False
++3163978416,Comparing SDTM and FHIR® for Real World Data from Electronic Health Records for Clinical Trial Submissions,['computational_biology'],computational_biology,['database_systems'],['database_systems'],True
++7135254915,Etude in silico des activités antioxydante et anti-inflammatoire des extraits des feuilles de Prunus persica et Rubus ulmefolius,['computational_biology'],computational_biology,[],['computational_biology'],False
++3174432239,KMDATA: a curated database of reconstructed individual patient-level data from 153 oncology clinical trials,['computational_biology'],computational_biology,['computational_biology'],['database_systems'],False
++4405278807,ПРОГНОЗИРОВАНИЕ БИОЛОГИЧЕСКОЙ АКТИВНОСТИ АРИЛИДЕНОВЫХ СОЕДИНЕНИЙ НА ОСНОВЕ РОДАНИНА И 2-ТИОБАРБИТУРОВОЙ КИСЛОТЫ,['computational_biology'],computational_biology,[],[],False
++7135069038,Perceptions of digital pathology among healthcare workers in pathology,['computational_biology'],computational_biology,['hci'],['hci'],False
++4402911702,Covariance Assisted Multivariate Penalized Additive Regression (CoMPAdRe),"['computational_biology', 'scientific_computing']",computational_biology,['ml_theory'],"['computational_biology', 'ml_theory']",False
++3159857369,Where Do We Stand in Regularization for Life Science Studies?,['computational_biology'],computational_biology,['ml_theory'],"['computational_biology', 'ml_theory']",False
++2786392319,India's technological climate: a harbinger in resolving and reducing the carbon footprint,"['robotics', 'sustainable_computing']",sustainable_computing,[],[],False
++3093665008,Managing Uncertainty in AI-Enabled Decision Making and Achieving Sustainability,['sustainable_computing'],sustainable_computing,['ai_safety'],['ai_agents'],True
++4387951271,A Comparative Analysis for Control of Heating and Cooling Systems of a Real Office Building with TRNSYS,['sustainable_computing'],sustainable_computing,['sustainable_computing'],[],False
++7168018191,From Nollywood to New Media: The Transformation of Cultural Narratives in Nigeria’s Digital Communication Landscape,['sustainable_computing'],sustainable_computing,[],[],False
++4410538484,The Comparative Analysis of the Effectiveness of the Energy Resources Using Fuzzy PROMETHE-TOPSIS Method,['sustainable_computing'],sustainable_computing,[],[],False
++4313647803,Interaction between extrinsic and intrinsic online review cues: perspectives from cue utilization theory,['sustainable_computing'],sustainable_computing,[],['hci'],False
++4404238308,Recent Advances and Future Challenges of Solar Power Generation Forecasting,['sustainable_computing'],sustainable_computing,[],[],False
++7215545478,PELATIHAN PEMANFAATAN E-LEARNING UNTUK PENGAJARAN DI KELAS PADA SMA MUHAMMADIYAH 1 PONOROGO,['sustainable_computing'],sustainable_computing,[],[],False
++4308155833,"Deriving gapless CO2 concentrations using a geographically weighted neural network: China, 2014–2020",['sustainable_computing'],sustainable_computing,['computer_vision'],[],False
++3080194848,First Insights on Soil Respiration Prediction across the Growth Stages of Rainfed Barley Based on Simulated MODIS and Sentinel-2 Spectral Indices,['sustainable_computing'],sustainable_computing,[],[],False
++4406505865,The Moderation Effect of Technology Anxiety on Digital Transformation Readiness in Public Universities: An Organizational Readiness Approach,['sustainable_computing'],sustainable_computing,[],[],False
++7141737583,AI Innovation in Business,['sustainable_computing'],sustainable_computing,[],[],False
++4408774499,"Role of AI Innovation, Clean Energy and Digital Economy towards Net Zero Emission in the United States: An ARDL Approach","['computational_complexity', 'distributed_computing', 'sustainable_computing']",sustainable_computing,[],[],False
++7220836233,Strategi Pengelolaan Teknologi Informasi dalam Meningkatkan Mutu Pendidikan Madrasah,['sustainable_computing'],sustainable_computing,[],[],False
++3001058942,Introduction: Role of Technical Reports in Sci-Tech Libraries,['sustainable_computing'],sustainable_computing,[],[],False
++2948997629,Sidewalk and Toronto: Critical Systems Heuristics and the Smart City,['sustainable_computing'],sustainable_computing,['hci'],"['hci', 'sustainable_computing']",True
++3025387569,Petit traité d'écologie numérique : rencontres nantaises,['sustainable_computing'],sustainable_computing,[],['sustainable_computing'],False
++4416214949,Electricity-Carbon Collaborative Optimization of Electrolytic Aluminum Load Considering Green Certificate-Carbon Trading and Demand Response,"['applied_cryptography', 'sustainable_computing']",sustainable_computing,[],[],False
++2905015752,Performance and power consumption evaluation in smartphone based image processing for medical applications,"['ai_safety', 'computer_architecture', 'computer_vision', 'hci', 'sustainable_computing']",sustainable_computing,['computer_architecture'],"['edge_computing', 'sustainable_computing']",False
++2950028967,Carbon Footprint Management: A Pathway Toward Smart Emission Abatement,['sustainable_computing'],sustainable_computing,[],[],False
++4392157044,"Information technology, gender economic inclusion and environment sustainability in sub-Sahara Africa",['sustainable_computing'],sustainable_computing,[],[],False
++4415684933,The Scale and Innovation Effects of Sci-Tech Finance Pilot Policy from the Perspective of Sustainable Development,['sustainable_computing'],sustainable_computing,[],[],False
++4323665427,Green Space Compactness and Configuration to Reduce Carbon Emissions from Energy Use in Buildings,['sustainable_computing'],sustainable_computing,[],[],False
++4399765326,The Influence of Environmental Factors on a Photovoltaic Panel–PVsyst Simulation,['sustainable_computing'],sustainable_computing,[],[],False
++4406559909,Efficient and adaptive design of RBF neural network for maximum energy harvesting from standalone PV system,"['scientific_computing', 'sustainable_computing']",sustainable_computing,['ai_agents'],[],False
++7117113370,The conservation outcome of Indo-Pacific humpback Dolphin in China: A habitat suitability-based assessment,['sustainable_computing'],sustainable_computing,['computational_biology'],[],False
++4390978249,Prdiction and Analysis on Provincial Carbon Emissions and Carbon Peak,['sustainable_computing'],sustainable_computing,[],[],False
++4361987249,Comparative study of the prediction of electrical energy from a photovoltaic system using the intelligent systems ANFIS and ANFIS-GA,['sustainable_computing'],sustainable_computing,[],[],False
++4399284355,Forecasting China's agricultural carbon emissions: A comparative study based on deep learning models,"['deep_learning_architectures', 'sustainable_computing']",sustainable_computing,[],[],False
++4390621478,"Raising Young Pupils and Students' Awareness Levels About the Impacts of Digital Technologies on the Environment, Climate, Health and Well-Being Through a Challenge",['sustainable_computing'],sustainable_computing,['hci'],['sustainable_computing'],False
++7163329089,Evaluation of Quipper LMS Utilization in the Junior High School Department as Basis for a Blended Learning Model Under the Revised MATATAG Curriculum,['quantum_computing'],quantum_computing,['edge_computing'],[],False
++7116076303,Quantum Radiometric Calibration,"['ml_theory', 'quantum_computing']",quantum_computing,['quantum_computing'],[],False
++2783787559,Consideration of analogies between magnetic and quantum notices for molecular network,"['applied_cryptography', 'computational_biology', 'quantum_computing']",quantum_computing,[],['quantum_computing'],False
++2964298217,Versatile Super-Sensitive Metrology Using Induced Coherence,['quantum_computing'],quantum_computing,[],[],False
++4411232250,Imperfect unambiguous state discrimination: practical case study for future networks,['quantum_computing'],quantum_computing,['computer_networks'],['quantum_computing'],False
++4411225084,Intrinsic quality factors approaching 10 million in superconducting planar resonators enabled by spiral geometry,"['quantum_computing', 'scientific_computing']",quantum_computing,[],['quantum_computing'],False
++2755062554,Measuring the Single-Photon Temporal-Spectral Wave Function,['quantum_computing'],quantum_computing,"['computer_architecture', 'quantum_computing']",[],True
++4390533496,Feature Stores: A Key Enabler for Feature Reusability and Availability Across Machine Learning Pipelines,"['cloud_computing', 'operating_systems', 'software_engineering']",operating_systems,['database_systems'],['database_systems'],False
++2964142687,A Read-leveling Data Distribution Scheme for Promoting Read Performance in SSDs with Deduplication,"['algorithms', 'database_systems', 'distributed_computing', 'operating_systems']",operating_systems,['computer_architecture'],"['computer_architecture', 'operating_systems']",False
++4400484472,Metodología innovadora para la optimización del guiado de visitantes en museos planetarios mediante aplicativos de realidad aumentada,['operating_systems'],operating_systems,['hci'],"['hci', 'software_engineering']",False
++4292057429,Robust private information retrieval for low encoding/decoding complexity distributed storage,"['applied_cryptography', 'computer_networks', 'cybersecurity', 'distributed_computing', 'hci', 'information_retrieval', 'operating_systems', 'privacy_preserving_computation']",operating_systems,"['computer_networks', 'information_retrieval']","['distributed_computing', 'privacy_preserving_computation']",False
++4254652521,Optical Storage Systems,['operating_systems'],operating_systems,['computer_architecture'],['computer_architecture'],False
++4225872280,Batteries Included: Ansible Modules,"['cloud_computing', 'cybersecurity', 'operating_systems', 'software_engineering']",operating_systems,[],[],False
++3189882660,VMware лицензии для учебных заведений | Прайс-лист,"['cloud_computing', 'operating_systems']",operating_systems,[],[],False
++4224674551,A Skew Logistic Distribution for Modelling COVID-19 Waves and Its Evaluation Using the Empirical Survival Jensen–Shannon Divergence,['operating_systems'],operating_systems,[],['ml_theory'],False
++4406461284,Comprehensive Charaterizations on Read Disturbs in QLC Charge-Trap (CT) 3D NAND Flash,['operating_systems'],operating_systems,['computer_architecture'],['computer_architecture'],False
++4293193165,A Frequency-Selective Digital Predistortion Method Based on a Generalized Indirect Learning Architecture,['operating_systems'],operating_systems,['computer_architecture'],[],False
++7117899816,Anatomy of Chaos: A Theoretical Framework for Forecasting the Morphology of Post-Crisis Regime (Vol. 1),['operating_systems'],operating_systems,[],[],False
++4380875641,Appliance Identification via Combinatorial Fusion Analysis-Assisted Bayesian-Optimized Classifier,['operating_systems'],operating_systems,"['database_systems', 'ml_theory']",[],False
++2973989512,Tiered Data Storage Model,"['cloud_computing', 'database_systems', 'distributed_computing', 'operating_systems']",operating_systems,['database_systems'],['operating_systems'],False
++4387067117,iOS Persistence Options,"['database_systems', 'operating_systems']",operating_systems,['software_engineering'],['software_engineering'],False
++3181207510,Development of an APP with IONIC - EDUAPP,['operating_systems'],operating_systems,"['hci', 'software_engineering']",['hci'],False
++2894822167,Testing and Improving Deployment of ATLAS Releases to CVMFS,"['cloud_computing', 'distributed_computing', 'operating_systems']",operating_systems,['operating_systems'],['distributed_computing'],False
++4401070015,Identifying the Relationship Structure Among Multiple Datasets Using Independent Vector Analysis: Application to Multi-Task fMRI Data,['operating_systems'],operating_systems,"['computer_vision', 'ml_theory']","['computational_biology', 'ml_theory']",False
++4392711967,Provisioning CSD-based Storage Systems with Erasure-coding Offloaded to the CSD,"['cloud_computing', 'database_systems', 'distributed_computing', 'operating_systems']",operating_systems,"['computer_architecture', 'database_systems']","['computer_architecture', 'operating_systems']",False
++3212917515,Effectiveness of Multi-Abstraction Computing Tools on Promoting Exploratory Self-learning in Engineering: a Case Study using a Custom Real-Time Operating System for Remote Learning,['operating_systems'],operating_systems,['hci'],['operating_systems'],False
++7137550668,CERN openlab summer students' lightning talks 1,"['cloud_computing', 'distributed_computing', 'operating_systems']",operating_systems,[],['distributed_computing'],False
++7213061804,SeqMoE: Toward Full-Load Performance via Predictive and Graph-Compatible MoE Offloading,"['cloud_computing', 'computer_architecture', 'database_systems', 'deep_learning_architectures', 'distributed_computing', 'operating_systems']",operating_systems,"['deep_learning_architectures', 'llm']","['ai_accelerators', 'llm']",False
++4408213108,Storage of Solid,"['operating_systems', 'software_engineering']",operating_systems,[],[],False
++3193637602,THE EMPLOYING OF MOBILE PHONES AND TABLETS APPLICATIONS IN PIANO TEACHING (A COMPARATIVE STUDY BETWEEN FLOWKEY AND YOUSICIAN),['operating_systems'],operating_systems,['hci'],['hci'],False
++2981636369,A Proposed Framework to Explore Semantic Relations for Learning Process Management,['operating_systems'],operating_systems,['hci'],[],False
++4200058389,Classical and objective Bayesian estimation and confidence intervals of an asymmetric loss based capability index Cpmk′$\mathcal {C}^{\prime }_{pmk}$,['operating_systems'],operating_systems,['ml_theory'],['ml_theory'],False
++3022441546,Data Science as a Route to AI for Middle- and High-School Students,['operating_systems'],operating_systems,['nlp'],[],False
++3113396090,Few Percent Efficient Polarization-Sensitive Conversion in Nonlinear Plasmonic Interactions Inside Oligomeric Gold Structures,"['operating_systems', 'scientific_computing']",operating_systems,[],[],False
++3038664825,Full Understanding of Hot Electrons and Hot/Cold Holes in the Degradation of p-channel Power LDMOS Transistors,['computer_architecture'],computer_architecture,['computer_architecture'],[],False
++4411019384,Can DOA Algorithms Be Deceived?,"['computer_architecture', 'scientific_computing']",computer_architecture,"['cybersecurity', 'ml_theory']",[],False
++3142700669,On the Automation of Periodic HardReal-time Processes: a Graph-Theoretical Approach,"['computer_architecture', 'operating_systems', 'programming_languages', 'robotics']",computer_architecture,"['algorithms', 'programming_languages']","['algorithms', 'formal_verification']",False
++4399426830,Analysis of Semiconductor Chip Performance using Medallion Architecture,"['applied_cryptography', 'computer_architecture', 'distributed_computing']",computer_architecture,['database_systems'],['database_systems'],False
++3082922165,Parameter Assignment And Schedulability Analysis For Real-Time Multiframe Task Systems,"['algorithms', 'computer_architecture', 'distributed_computing', 'formal_verification', 'operating_systems', 'robotics']",computer_architecture,['distributed_computing'],"['algorithms', 'operating_systems']",False
++2888777304,Large Room Temperature Charge-to-Spin Conversion Efficiency in Topological Insulator/CoFeB bilayers,['computer_architecture'],computer_architecture,['computer_architecture'],[],False
++4416146873,Reproducibility Report for SC25 Paper Story of Two GPUs: Characterizing the Resilience of Hopper H100 and Ampere A100 GPUs,"['ai_accelerators', 'computer_architecture', 'distributed_computing']",computer_architecture,['computer_architecture'],['distributed_computing'],False
++3004719967,Assessment of contribution of human factor and factors of material-technical supply to safety risks due to poor repair and technology,['computer_architecture'],computer_architecture,[],[],False
++4411122680,APPLICATION OF SIMULATION MODELLING FOR THE ANALYSIS OF TONAL RAIL TRACK CIRCUIT OPERABILITY,['scientific_computing'],scientific_computing,['computer_architecture'],[],False
++4415164777,Noise-Tolerant Hybrid Approach for Data-Driven Predictive Control,['scientific_computing'],scientific_computing,['ml_theory'],[],False
++6940648869,An analysis of an LQR design for a hybrid power plant’s load-sharing problem,['scientific_computing'],scientific_computing,[],[],True
++3134502016,A research parasite's perspective on establishing a baseline to avoid errors in secondary analyses,['scientific_computing'],scientific_computing,[],[],False
++4406816254,Corrective Lateral Transshipment Application in a Centralized Inventory System With Random Demand,['scientific_computing'],scientific_computing,['scientific_computing'],[],False
++4206279532,"Back Propagation Neural Network with Feature Sensitivity Analysis in Railways: Short Term Prediction of Passenger Flow using Time, Environmental and Operational Factors",['scientific_computing'],scientific_computing,['nlp'],[],False
++3210446440,Towards Improved Robustness of Public Transport by a Machine-Learned Oracle,['scientific_computing'],scientific_computing,"['algorithms', 'information_retrieval']",['algorithms'],False
++4405077071,Research on system parameter optimisation in electromagnetic tomography technology,['scientific_computing'],scientific_computing,['computer_vision'],['scientific_computing'],False
++4291941306,Lie Symmetry Analysis of the One-Dimensional Saint-Venant-Exner Model,['scientific_computing'],scientific_computing,[],[],False
++3216320654,Optimal Tracking Control for Unknown Linear Systems with Finite-Time Parameter Estimation,['scientific_computing'],scientific_computing,['ml_theory'],[],False
++4408094557,Nonlinear modelling of transient tyre dynamics,['scientific_computing'],scientific_computing,[],[],False
++2963395393,Homogenization and boundary layers in domains of finite type,['scientific_computing'],scientific_computing,[],[],False
++4382681213,A new family of Sakurai–Torii–Sugiura type iterative methods with high order of convergence,['scientific_computing'],scientific_computing,[],['scientific_computing'],False
++3082621528,Design of Bio-Impedance Electrode Topologies for Specific Depth Sensing in Skin Layer,['scientific_computing'],scientific_computing,['computer_architecture'],['scientific_computing'],False
++3157604584,Aggregation and disaggregation of active particles on the unit sphere with time-dependent frequencies,['scientific_computing'],scientific_computing,['scientific_computing'],[],False
++4394993168,A mathematical analysis of the corruption dynamics model with optimal control strategy,['scientific_computing'],scientific_computing,[],[],False
++2935049866,Frequency Limited Model Reduction Techniques for Discrete-Time Systems,['scientific_computing'],scientific_computing,['ml_theory'],['scientific_computing'],False
++3000672483,Stochastic quantization approach to modeling of constrained Ito processes in discrete time,['scientific_computing'],scientific_computing,['ml_theory'],['ml_theory'],True
++3088835008,Beidou-GPS dual mode positioning of mobile communication equipment based on Android platform,['scientific_computing'],scientific_computing,['computer_networks'],['computer_networks'],False
++4322394327,"Interfacing Power System and ICT Simulators: Challenges, State-of-the-Art, and Case Studies",['scientific_computing'],scientific_computing,['computer_networks'],"['computer_networks', 'scientific_computing']",False
++3123748159,The diversity bonus in pooling local knowledge about complex problems,['scientific_computing'],scientific_computing,[],[],False
++4406141980,Elliptic Neumann problems with highly discontinuous nonlinearities,['scientific_computing'],scientific_computing,[],[],False
++2810988304,Optimal Stopping of McKean--Vlasov Diffusions via Regression on Particle Systems,['scientific_computing'],scientific_computing,['algorithms'],['scientific_computing'],False
++2977976693,Determination of luminance distribution under tropical sky conditions,['federated_learning'],federated_learning,[],[],False
++4401880100,A Decentralised Asynchronous Optimisation Algorithm with an Application to Phase Retrieval,['federated_learning'],federated_learning,"['distributed_computing', 'robotics']","['algorithms', 'distributed_computing']",False
++4385602851,The Ionospheric Three-Dimensional Electron Density Variations Induced by the 21 August 2017 Total Solar Eclipse by Using Global Ionospheric Specification,['federated_learning'],federated_learning,[],[],False
++2921727985,Entrance Pupil Irradiance Estimating Model for a Moon-Based Earth Radiation Observatory Instrument,['federated_learning'],federated_learning,[],[],False
++4407950319,Enabling Computer Science Students to Choose Their Own Adventure in a Mandatory and Curricular Challenge-Based Learning Course,['federated_learning'],federated_learning,[],[],False
++7148322461,Predictability can be dynamically constructed in deterministic systems,['federated_learning'],federated_learning,['ml_theory'],[],False
++3215093560,A Fine PolSAR Terrain Classification Algorithm Using the Texture Feature Fusion-Based Improved Convolutional Autoencoder,['federated_learning'],federated_learning,"['computer_vision', 'deep_learning_architectures']","['computer_vision', 'deep_learning_architectures']",False
++7163589765,The Impact of AI ‐Supported Think‐Pair‐Share Instruction on Students' Problem‐Solving Skills and Motivation in Learning Quadratic Equations: A Social Constructivist Perspective,['federated_learning'],federated_learning,"['ai_agents', 'hci']",['hci'],True
++7204715907,EFFECTS OF ARTIFICIAL INTELLIGENCE (AI) ON STUDENTS' CRITICAL THINKING TOWARDS ACADEMIC ACHIEVEMENT IN MODIBBO ADAMA UNIVERSITY YOLA (MAU),['federated_learning'],federated_learning,['hci'],[],False
++2907334421,Basic Study of Solar Battery Powered Wireless Power Transfer System with MPPT Mode and DC Bus Stabilization for Lunar Rover,['federated_learning'],federated_learning,['computer_architecture'],[],True
++2908566518,Epochs of an AI cosmology,"['ai_safety', 'federated_learning']",federated_learning,[],[],False
++7143299875,Efektivitas Penyimpanan Google Drive dalam Menyimpan Data,['federated_learning'],federated_learning,[],[],False
++4254855164,Artificial Intelligence,"['ai_safety', 'federated_learning']",federated_learning,[],[],False
++4402145393,Towards practical artificial intelligence in Earth sciences,"['ai_safety', 'foundation_models']",foundation_models,['scientific_computing'],['foundation_models'],False
++7171486454,Neonatal Hypoxic-ischaemic Encephalopathy Classification from the EEG and HRV Signals Using a Conformer based Masked Autoencoder,"['deep_learning_architectures', 'foundation_models', 'ml_theory']",foundation_models,"['computational_biology', 'deep_learning_architectures']","['computational_biology', 'foundation_models']",False
++4416183290,GLOF-Vulnerability Identification Based on Automatic Method for Mapping Glacial Lakes in Cordillera Blanca (Peru),"['cybersecurity', 'foundation_models']",foundation_models,"['computer_vision', 'scientific_computing']",['computer_vision'],False
++4412605435,Multi-Source Joint Adaptive Distribution With Online Transfer Learning for Cross-Domain Fault Diagnosis,['foundation_models'],foundation_models,['deep_learning_architectures'],['deep_learning_architectures'],False
++7172007151,A meta-learning method with reduced data requirements for training and updating deep learning models on wearable devices,"['deep_learning_architectures', 'foundation_models']",foundation_models,"['deep_learning_architectures', 'ml_theory']","['computational_biology', 'edge_computing']",False
++7171698092,A Distributional Robustness Margin For Pathology Foundation Models,"['computational_biology', 'computer_vision', 'foundation_models']",foundation_models,"['ai_safety', 'computer_vision']","['ai_safety', 'foundation_models']",False
++4412420015,Few-shot partial multi-label learning with credible non-candidate label,"['foundation_models', 'nlp']",foundation_models,['ml_theory'],['ml_theory'],False
++4385757412,Semi-Supervised Learning for Low-Cost Personalized Obstructive Sleep Apnea Detection Using Unsupervised Deep Learning and Single-Lead Electrocardiogram,"['database_systems', 'deep_learning_architectures', 'foundation_models']",foundation_models,['computational_biology'],['computational_biology'],False
++3160088594,Few-Shot Learning for Ct Scan Based Covid-19 Diagnosis,"['deep_learning_architectures', 'foundation_models']",foundation_models,['computer_vision'],['computer_vision'],False
++4313024900,Unsupervised Domain Adaptation across FMCW Radar Configurations Using Margin Disparity Discrepancy,"['computer_vision', 'deep_learning_architectures', 'foundation_models', 'ml_theory']",foundation_models,"['computer_architecture', 'deep_learning_architectures']",['ml_theory'],False
++4402355991,Self‐supervised representation learning of metro interior noise based on variational autoencoder and deep embedding clustering,"['deep_learning_architectures', 'foundation_models']",foundation_models,"['computer_vision', 'speech_audio']",['deep_learning_architectures'],False
++4413751397,Deep insights into cognitive decline: A survey of leveraging non-intrusive modalities with deep learning techniques,"['deep_learning_architectures', 'foundation_models']",foundation_models,"['computational_biology', 'computer_vision', 'deep_learning_architectures', 'foundation_models', 'llm', 'multimodal_ai', 'nlp', 'speech_audio']","['computational_biology', 'multimodal_ai', 'speech_audio']",False
++3139784666,Transfer Learning in Smart Environments,"['algorithms', 'foundation_models', 'ml_theory']",foundation_models,['ml_theory'],['edge_computing'],False
++4391216411,Capacity optimal allocation of wind-solar-fire low-carbon power generation system based on green certificate-carbon emission equivalent interaction mechanism,['applied_cryptography'],applied_cryptography,['sustainable_computing'],[],False
++4401369144,A dynamic model of repositioning with a Markov-switching regime,"['applied_cryptography', 'scientific_computing']",applied_cryptography,[],[],False
++3011920068,Reliability Enhancement for Multi-level Cell NAND Flash Memory Using Error Asymmetry,"['applied_cryptography', 'computer_networks', 'operating_systems']",applied_cryptography,['computer_architecture'],['computer_architecture'],False
++3126856290,Multi-Stage Coordination Volt/VAR Control with CVR in Active Distribution Network in presence of Inverter-Based DG Units and Soft Open Points,"['applied_cryptography', 'scientific_computing']",applied_cryptography,"['computer_architecture', 'scientific_computing']",[],False
++4226498265,Knowledge-based optimal irrigation scheduling of agro-hydrological systems,['applied_cryptography'],applied_cryptography,"['ml_theory', 'scientific_computing']",[],False
++4405179471,"The Polish Cryptographers Marian Rejewski, Jerzy Różycki, and Henryk Zygalski",['applied_cryptography'],applied_cryptography,[],[],False
++4391590938,Spin: An Efficient Secure Computation Framework with GPU Acceleration,"['ai_accelerators', 'applied_cryptography', 'computer_networks', 'cybersecurity', 'deep_learning_architectures', 'distributed_computing', 'federated_learning', 'ml_theory', 'privacy_preserving_computation']",applied_cryptography,"['applied_cryptography', 'deep_learning_architectures', 'distributed_computing']","['ai_accelerators', 'privacy_preserving_computation']",False
++4378805032,A Frequency Control Strategy for EV Stations Based on MPC-VSG in Islanded Microgrids,['applied_cryptography'],applied_cryptography,['computer_architecture'],[],True
++3090466325,Finite field elements of high order on a base of modified Gao approach.,['applied_cryptography'],applied_cryptography,['algorithms'],"['algorithms', 'applied_cryptography']",False
++4232616731,Data Visualization and Information design: bringing data to life,['computer_graphics'],computer_graphics,['hci'],['computer_graphics'],False
++3035417602,Perancangan iklan layanan masyarakat tentang pencegahan bunuh diri akibat depresi pada usia dewasa awal melalui media audio visual,['computer_graphics'],computer_graphics,[],[],False
++2785914471,Assessing clutter reduction in parallel coordinates using image processing techniques,"['computer_graphics', 'computer_vision']",computer_graphics,"['computer_vision', 'hci']",['computer_graphics'],False
++4414732984,Chord-Color Mapping for Audio-Visual Generation: Integrating Deep Learning and Emotion Quantification,"['computer_graphics', 'deep_learning_architectures']",computer_graphics,"['generative_ai', 'multimodal_ai', 'speech_audio']","['generative_ai', 'multimodal_ai', 'speech_audio']",False
++4389429756,#peaches,"['ai_safety', 'computer_graphics']",computer_graphics,[],[],False
++4229820775,Game of X V.2,['computer_graphics'],computer_graphics,[],[],False
++4409753672,PERANCANGAN IDENTITAS VISUAL DALAM MENINGKATKAN BRAND RECOGNITION PEKEN SENGGOL GURAT,['computer_graphics'],computer_graphics,[],[],False
++7163058269,Interactive Strategies for Coordinated Exploration of Timeline and Directly-Follows Graph Visualizations,['computer_graphics'],computer_graphics,"['hci', 'information_retrieval']","['computer_graphics', 'hci']",False
++4404698773,AN APPLIED RESEARCH ON THE USE OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES IN MOVING IMAGE PRODUCTION,['computer_graphics'],computer_graphics,"['generative_ai', 'hci']",['generative_ai'],False
++7206470547,Modelado de datos en el plano,['computer_graphics'],computer_graphics,[],[],True
++4375835151,Video Animasi 3D Cluster The Rich Sebagai Media Promosi Pada Ciptaland Development,['computer_graphics'],computer_graphics,[],['computer_graphics'],False
++4392521761,PENGEMBANGAN ANIMASI 3D VENTILATOR PADA STEQHOQ ROBOTIKA INDONESIA,['computer_graphics'],computer_graphics,[],['computer_graphics'],False
++2911494573,BASMAT,"['compilers', 'computer_architecture', 'computer_graphics', 'distributed_computing', 'scientific_computing']",computer_graphics,"['computer_architecture', 'distributed_computing']","['ai_accelerators', 'scientific_computing']",False
++2804059852,Local blur analysis and phase error correction method for fringe projection profilometry systems,['computer_graphics'],computer_graphics,['computer_vision'],['computer_vision'],False
++3157008771,Ensemble methods for ranking data with and without position weights,['multimodal_ai'],multimodal_ai,"['information_retrieval', 'ml_theory']",['ml_theory'],False
++7220827782,Analisis Performa Summary-Based Multimodal Retrieval dengan Arsitektur Dual Database pada Sistem RAG Asuhan Gizi Anak,"['computer_vision', 'information_retrieval', 'llm', 'multimodal_ai']",multimodal_ai,[],"['information_retrieval', 'multimodal_ai']",False
++7197818024,Метод та програмні засоби мультимодального аналізу медичних даних на основі глибокого навчання,['multimodal_ai'],multimodal_ai,"['computer_vision', 'deep_learning_architectures']","['computational_biology', 'computer_vision']",False
++7118751186,Digital Library Services in Arts and Sciences Colleges of Chennai District - A Study,['information_retrieval'],information_retrieval,[],[],False
++7135293231,Three Essays on Paid Q&A: Understanding Content Consumption and Social Engagement,"['hci', 'information_retrieval']",information_retrieval,[],['hci'],False
++7204876499,Visible but Not Yet Curatable: Characterizing the Curatability of Compact and Derived Open LLM Artifacts,"['autonomous_scientific_discovery', 'hci', 'information_retrieval', 'llm', 'software_engineering']",information_retrieval,"['digital_libraries_placeholder_none', 'llm']",['information_retrieval'],False
++3175543912,„Exploratory search” – kłopoty architekta informacji z terminologią,"['hci', 'information_retrieval']",information_retrieval,[],"['hci', 'information_retrieval']",False
++2798885959,Assessing the Readability of Web Search Results for Searchers with Dyslexia,"['information_retrieval', 'nlp']",information_retrieval,['hci'],"['hci', 'information_retrieval']",False
++7206221668,Інформаційний бюлетень НТБ НАУ №12(2019),['information_retrieval'],information_retrieval,[],[],True
++4399261954,Web scraping Based Site For Product Analysis,['information_retrieval'],information_retrieval,['information_retrieval'],[],True
++7221540348,A child recognition system based on image selection patterns: Work in progress,"['cybersecurity', 'information_retrieval']",information_retrieval,"['cybersecurity', 'hci']",['hci'],False
++3208034694,Automated Evaluation of Web Site Accessibility Using A Dynamic Accessibility Measurement Crawler,"['information_retrieval', 'software_engineering']",information_retrieval,"['cybersecurity', 'hci']","['hci', 'software_engineering']",False
++7206653269,Інформаційний бюлетень НТБ НАУ № 10 (2022),['information_retrieval'],information_retrieval,[],[],False
++2801495198,Expert Internet Searching (fifth edition). (2017),"['computational_complexity', 'information_retrieval']",information_retrieval,[],[],False
++4389683448,The Effect of the Personalized Recommendation System of Online Shopping Platform on Consumers’ Purchase Intention,['information_retrieval'],information_retrieval,"['hci', 'information_retrieval']",['hci'],True
++6910675515,Towards Industrial Convergence : Understanding the evolution of scientific norms and practices in the field of AI,"['information_retrieval', 'scientific_computing']",information_retrieval,[],[],False
++7220856794,The Potential Influence of Recommendation Algorithms on Personal Self-Narrative Construction : A Narrative Identity-Based Literature Review Centered on YouTube and Shorts with Christian Reflection,['information_retrieval'],information_retrieval,[],[],False
++4412395283,Physical activity and diet just-in-time adaptive intervention to reduce blood pressure: a randomized controlled trial,"['compilers', 'computer_architecture']",compilers,[],[],False
++3179712771,Atomic grammar: a learning source of ESL error identification and correction / Nor Syamimi Iliani Che Hassan … [et al.],['compilers'],compilers,[],[],False
++7213884990,PENINGKATAN LITERASI BAGI SISWA SMA MELALUI PELATIHAN PENYUSUNAN LITERATURE REVIEW,['compilers'],compilers,[],[],False
++7221269676,RANCANGAN PROGRAM OTOMASI UNTUK DESAIN SINGLE LINE DIAGRAM (SLD) PADA INSTALASI LISTRIK BERBASIS AUTOLISP DAN STANDAR PUIL 2020,"['compilers', 'database_systems']",compilers,['software_engineering'],[],False
++2591985246,Integration Of C Into An Introductory Course In Machine Organization,"['compilers', 'operating_systems']",compilers,['programming_languages'],[],False
++4282571853,SPEED CONTROL OF DC MOTOR USING PID AND FUZZY LOGIC CONTROLLER,['compilers'],compilers,['ml_theory'],[],False
++4410182237,Implementation method forward chaining in game puzzle (case study in Paud Dini Laras Yogyakarta),['compilers'],compilers,['hci'],[],False
++2908836625,Optimisation de la compilation de règles métier probabilistes à l'aide de PRM,['compilers'],compilers,['database_systems'],['compilers'],False
++4401497158,A hybrid inductor‐current sensor with fast response and high accuracy using online calibration for high‐frequency BUCK DC‐DC converters,['compilers'],compilers,['computer_architecture'],['computer_architecture'],False
++2992990171,Research of College English Teaching Methods Based on New Engineering Talents Training,['compilers'],compilers,['nlp'],[],False
++2794424394,H∞ control of memristive neural networks with aperiodic sampling and actuator saturation,"['ai_accelerators', 'compilers']",compilers,['ml_theory'],[],False
++4206295199,Reduced-Space Relevance Vector Machine for Adaptive Electrical Capacitance Volume Tomography,"['algorithms', 'compilers', 'computational_complexity', 'quantum_computing']",compilers,"['computer_vision', 'ml_theory']",['ml_theory'],False
++4403183224,Stability Analysis and Stabilization of Semi-Markov Jump Linear Systems With Improved Efficiency of Probabilistic Information Utilization,['compilers'],compilers,[],[],False
++3134130217,Recursos educativos digitales como herramienta didáctica en el subnivel inicial II de la Unidad Educativa Alemán,['compilers'],compilers,[],[],False
++2886921199,Wind Turbine Aerodynamic Load Fluctuation Reduction Using Model Based Iterative Learning Control,"['compilers', 'scientific_computing']",compilers,['ml_theory'],[],False
++4226087031,Hands-On Computer Science: The Array of Things Experimental Urban Instrument,"['compilers', 'computer_networks']",compilers,['computer_networks'],"['edge_computing', 'hci']",False
++4392445319,Discovering Players’ Problem-Solving Behavioral Characteristics in a Puzzle Game through Sequence Mining,['compilers'],compilers,['hci'],['hci'],False
++4416730187,PENGEMBANGAN APLIKASI SIAGUD UNTUK PENINGKATAN KINERJA GURU PAI DI MAN 2 PADANGSIDIMPUAN,['compilers'],compilers,[],['software_engineering'],True
++7142973430,Development Of Digital Literacy And Environmental Literacy Instruments In Ecology And Biodiversity Materials,['compilers'],compilers,['software_engineering'],[],False
++3161777593,"The Construction of New Type Loose Leaf Textbook in Vocational Education under the Concept of the Reform of Teachers, Teaching Materials and Teaching Methods",['compilers'],compilers,[],[],False
++4290851516,Selection of unlabeled source domains for domain adaptation in remote sensing,"['compilers', 'programming_languages']",compilers,['ml_theory'],['computer_vision'],False
++4388462287,Microphone Use Techniques,['speech_audio'],speech_audio,[],['speech_audio'],True
++4307835756,Passive localisation of mixed near‐field and far‐field signals using the transformed symmetric nested array,['speech_audio'],speech_audio,"['algorithms', 'speech_audio']",[],False
++4387489607,A DOA estimation method for array signals based on deep learning,"['algorithms', 'speech_audio']",speech_audio,"['deep_learning_architectures', 'speech_audio']",['deep_learning_architectures'],False
++4412729123,Building the Raw Balance,['speech_audio'],speech_audio,[],[],True
++7172368152,Insights and Applications in Piano Performance through Acoustics and Mechanics,['speech_audio'],speech_audio,[],[],False
++2965346518,Some Approaches To Representing Sound With Colour And Shape,"['applied_cryptography', 'computer_graphics', 'computer_vision', 'speech_audio']",speech_audio,['hci'],['hci'],False
++4212954326,Deconvolved Conventional Beamforming and Uncertainty Analysis for a Difference Coarray,"['computer_vision', 'speech_audio']",speech_audio,['computer_architecture'],[],False
++3134129880,"‘Smart speaker, tell me about your acoustic sensor’",['speech_audio'],speech_audio,['hci'],[],False
++2991196924,Intégrer les LOD dans un cube de données : Transformons une action technique en valeur organisationnelle,['database_systems'],database_systems,['database_systems'],[],False
++4307637929,The Next Frontier in Intelligent Augmentation: Human-Machine Collaboration in Strategic Marketing Decision-Making,['database_systems'],database_systems,['hci'],['hci'],False
++2920557606,Performance analysis of a hyperspectral system for human iris imaging,['database_systems'],database_systems,['computer_vision'],['computer_vision'],False
++7131351419,Construction and Effectiveness Evaluation of an Artificial Intelligence-Assisted Student Affairs Processing Model for Counselors,['database_systems'],database_systems,['hci'],['hci'],False
++7159658581,Clasificador Naive como herramienta de decisión en la contratación de personal académico del TESCo,['database_systems'],database_systems,['ml_theory'],[],False
++4401268941,Data Science for Librarians,['database_systems'],database_systems,[],[],False
++4387233410,Prediction of Student Graduation with the Neural Network Method Based on Particle Swarm Optimization,['database_systems'],database_systems,['ml_theory'],[],False
++2997642018,Aplikasi Pendataan Dan Pengolahan Gaji Di Cv Master Laundry,"['database_systems', 'software_engineering']",database_systems,[],['software_engineering'],False
++7118687013,Exploring the Role of Generative AI and Deep Learning in Personalized Education Content: A Systematic Literature Review,"['ai_safety', 'algorithms', 'database_systems', 'generative_ai', 'hci', 'privacy_preserving_computation']",database_systems,"['llm', 'nlp']","['generative_ai', 'hci']",False
++4401822397,Prediksi Kelulusan Tepat Waktu Berdasarkan Riwayat Akademik Menggunakan Metode K-Nearest Neighbor,['database_systems'],database_systems,['ml_theory'],[],False
++3201407754,"Prediksi Prestasi Siswa SMK Masuk Pasar Kerja Menggunakan Teknik Data Mining (Studi Kasus SMKN 1 KotaAgung Timur Tanggamus, Lampung)","['database_systems', 'ml_theory']",database_systems,['ml_theory'],[],False
++4205179238,SISTEM INFORMASI PEMINJAMAN BUKU DI PERPUSTAKAAN MADRASAH IBTIDAIYAH AL-MADINAH,['database_systems'],database_systems,['hci'],['software_engineering'],False
++4416374389,"Sistema web para la optimización de los procesos de ventas e inventario de la empresa Alis, Lima - 2023",['database_systems'],database_systems,['software_engineering'],['software_engineering'],False
++3036960336,Near Real Time Feedback of Seasonal Influenza Vaccination and Virological Sampling: Dashboard Utilisation in a Primary Care Sentinel Network,['database_systems'],database_systems,['hci'],['computational_biology'],False
++3189460411,The synergistic effect of ego-network stability and whole network position: a perspective of transnational coopetition network,"['computer_networks', 'database_systems']",database_systems,[],[],False
++3022939691,An Approach for Complex Event Streams Processing and Forecasting,"['database_systems', 'information_retrieval']",database_systems,['database_systems'],['ml_theory'],False
++3043717349,Toward Universal Data Interoperability in Networked Belief Models,"['computer_networks', 'database_systems', 'edge_computing']",database_systems,['ml_theory'],['ml_theory'],False
++7212066750,การพฒนาระบบบรหารจดการฟตเนสบนเวบสำหรบศนยฟตเนส SU.ED,"['ai_agents', 'cloud_computing', 'database_systems', 'software_engineering']",database_systems,['hci'],['software_engineering'],False
++7128354711,Design Get Up System Information Inventory Goods Based Web Using the Waterfall Method at CV. Tekno Maju Jaya,"['database_systems', 'information_retrieval', 'programming_languages']",programming_languages,"['database_systems', 'software_engineering']",['software_engineering'],False
++3091828950,A research model in didactics of programming,"['algorithms', 'programming_languages']",programming_languages,[],[],False
++4206497465,"Ownership, Moves, and Copies",['programming_languages'],programming_languages,[],['programming_languages'],False
++7140223191,Instruction Set and Language for Symbolic Regression,"['autonomous_scientific_discovery', 'computational_biology', 'computer_architecture', 'nlp', 'programming_languages']",programming_languages,"['algorithms', 'information_retrieval']",['algorithms'],False
++3204449783,An Evaluation Study of an E-Learning Course at the Duhok Polytechnic University,['programming_languages'],programming_languages,['hci'],[],False
++7220384869,PENGEMBANGAN E-COMMERCE USAHA KECIL MENENGAH KAMPUNG BATIK BETUNGAN BERBASIS WEB,"['database_systems', 'programming_languages']",programming_languages,['software_engineering'],['software_engineering'],False
++2892794698,Web and Android Programming Course Information System,['programming_languages'],programming_languages,['software_engineering'],['software_engineering'],False
++4413180303,Power Flow Visualization and Prediction of Perturbations in Smart Grids Using a Graphical Model,['programming_languages'],programming_languages,['scientific_computing'],[],False
++2990464075,Plenary sessions Plenary talk I: “The Origin and Goal of Future in CPSS: Industries 4.0 and Industries 5.0”,"['computer_architecture', 'distributed_computing', 'edge_computing', 'programming_languages', 'scientific_computing']",programming_languages,[],[],False
++3199165787,Development of Web-Based Tracer Alumni Information System,"['database_systems', 'programming_languages']",programming_languages,[],['software_engineering'],False
++3010153186,Linear Likelihood Approximation Filter for Nonlinear State Estimation,['programming_languages'],programming_languages,['ml_theory'],['ml_theory'],False
++3030633439,Research on the Value of Smarter Education in the Era of Big Data,"['cloud_computing', 'cybersecurity', 'programming_languages']",programming_languages,[],[],False
++2941342584,Integrating Smartphone with Study Activities of Universitas Muhammadiyah Surakarta,['programming_languages'],programming_languages,['hci'],['software_engineering'],False
++7215646023,Implementasi Aplikasi Sistem “Pakar Anggrek” Berbasis Android Di Kebun Raya Liwa,['programming_languages'],programming_languages,"['computational_biology', 'cybersecurity']",[],False
++4287283171,An Epistemic Separation Logic with Action Models,"['ai_agents', 'formal_verification', 'programming_languages']",programming_languages,['ml_theory'],"['formal_verification', 'programming_languages']",False
++4388795790,SOSIALISASI PENGENALAN APP INVENTOR BERBASIS MOBILE PROGRAMMING,"['hci', 'programming_languages']",programming_languages,[],[],False
++4402990505,A Study of Impact of Digital Technology and Use of Blockchain Technology from the Consumer Point of View,"['blockchain', 'computer_networks', 'distributed_computing']",blockchain,[],['blockchain'],False
++2977293541,Crashkurs Blockchain- inkl. Arbeitshilfen online,['blockchain'],blockchain,[],['blockchain'],False
++7214298050,Is America Ready for Crypto or Crypt-No: Understanding Cryptocurrency User’s Willingness to Federal Regulations,"['blockchain', 'cybersecurity']",blockchain,[],[],False
++3014880043,DBSCAN Algoritması Kullanarak Bitcoin Fiyatlarında Anormallik Tespiti,['blockchain'],blockchain,['database_systems'],[],False
++4416430007,Digital Product Passport Data Transfer via Dataspace Connector: Technical Implementation and Analysis,"['blockchain', 'formal_verification', 'privacy_preserving_computation']",blockchain,['blockchain'],['distributed_computing'],True
++3194765017,Elon Musk selling techno track he produced about NFTs as an NFT,['blockchain'],blockchain,[],[],False
++4382770016,The Effects of Blockchain Technology on Corporate Governance: Evidence from Emerging Economy,"['blockchain', 'database_systems']",blockchain,['blockchain'],[],False
++4245301215,Decentralization,['blockchain'],blockchain,[],['blockchain'],False
++3143290424,A Study on the Blockchain in Business ：Research Trends and Issues between Korea and Overseas,['blockchain'],blockchain,[],['blockchain'],False
++7117461924,The impact mechanism of digital economy platforms supported by blockchain and artificial intelligence on consumer green payment behavior,['blockchain'],blockchain,[],['blockchain'],False
++4417496166,"Blockchain Adoption, Consumer Trust, and Loyalty in Retail SMEs: Evidence from Southeast Nigeria","['blockchain', 'database_systems']",blockchain,[],['blockchain'],False
++4411360648,Research Trends in Blockchain Studies in South Korea Using Topic Modeling and Networks Analysis,"['blockchain', 'nlp']",blockchain,[],['blockchain'],False
++4386500200,Percepcja współczesnego pieniądza i kryptowalut w świetle badań ankietowych na przykładzie młodych użytkowników pieniądza z Czech i Polski,['blockchain'],blockchain,[],[],False
++4210686246,"Digital currencies: Challenges between the United States, the European Union and Asia Pacific",['blockchain'],blockchain,[],[],False
++4285316846,The Impact of Crypto-Asset Utilization as Payment Instrument toward Rupiah as Legal Tender in Indonesia,['blockchain'],blockchain,[],[],True
++3198186069,APRESENTAÇÃO DO DOSSIÊ,['blockchain'],blockchain,[],[],True
++4385250044,Design and Construction of Carbon Accounting Techniques in FinTech: A Scientometric Analysis,"['blockchain', 'cloud_computing', 'database_systems', 'sustainable_computing']",blockchain,['blockchain'],[],False
++7123350921,Arsitektur Unified Ledger untuk Donasi Digital yang Transparan dengan Idempotency dan Server-Sent Events,"['blockchain', 'distributed_computing']",blockchain,"['blockchain', 'database_systems']","['database_systems', 'software_engineering']",False
++3010373421,Joining Forces: Understanding Organizational Roles in Inter-organizational Smart Service Systems Engineering,"['cloud_computing', 'computer_networks']",cloud_computing,['hci'],['software_engineering'],False
++4391093673,Research on Computer Big Data Industrial Structure Adaptation Simulation System,['cloud_computing'],cloud_computing,"['database_systems', 'distributed_computing']","['cloud_computing', 'database_systems']",False
++4407956345,Evaluation and Analysis of University Education Quality Based on Internet Artificial Intelligence Big Data Algorithm,"['cloud_computing', 'computer_networks']",cloud_computing,[],[],False
++4248407127,Datenbereiche und Datendateien,"['cloud_computing', 'distributed_computing', 'operating_systems']",cloud_computing,[],[],False
++3094723118,Self-Learning Threshold-Based Load Balancing,"['cloud_computing', 'computer_networks', 'distributed_computing', 'scientific_computing']",cloud_computing,"['distributed_computing', 'ml_theory', 'reinforcement_learning']","['algorithms', 'cloud_computing']",False
++4389723967,Application Research of Computer Artificial Intelligence Technology in Green Cold Chain Logistics of Agricultural Products,"['cloud_computing', 'operating_systems']",cloud_computing,['cloud_computing'],[],False
++7110667317,GenAI alati u nastavi povijesti,['generative_ai'],generative_ai,['llm'],['hci'],True
++7166546311,Hybrid Modeling to Predict the Impact of Generative AI on Students' Learning Retention,"['generative_ai', 'ml_theory', 'scientific_computing']",generative_ai,['hci'],['hci'],False
++4409570677,Λ CDM and early dark energy in latent space: A data-driven parametrization of the CMB temperature power spectrum,['generative_ai'],generative_ai,"['ml_theory', 'scientific_computing']",['scientific_computing'],False
++4413402631,Generative AI in Multicultural Classrooms,"['ai_safety', 'generative_ai']",generative_ai,['hci'],['hci'],False
++7155650775,Generative AI as a Mediational Agent: Rethinking Learning in Sociocultural Theory,['generative_ai'],generative_ai,[],['hci'],False
++4414361486,From perception to policy: course coordinators’ views on generative artificial intelligence in higher education,['generative_ai'],generative_ai,[],[],False
++7165127855,Epistemic Cultures and AI Adoption in History Education: A Conceptual Framework with an Exploratory Czech Case Study,['generative_ai'],generative_ai,[],[],False
++7220858792,Research on AIGC-Enabled Human-Machine Collaborative Teaching Design Optimization for Digital Trade Micro-Major Curriculum Systems in Universities,"['generative_ai', 'privacy_preserving_computation']",generative_ai,[],[],False
++7204970079,From GenAI Virtual Patient Dialogue Logs to Teacher-Interpretable Process Evidence: A Learning Analytics Study in Higher Education,"['generative_ai', 'hci', 'nlp']",generative_ai,['llm'],['hci'],False
++4381744251,A Collaborative Learning Method based on Cross-domain Representation Correlation Congruence Distillation,"['computer_vision', 'deep_learning_architectures', 'foundation_models', 'generative_ai']",generative_ai,['nlp'],"['computer_vision', 'deep_learning_architectures']",False
++7113405664,TYCOV: Human Detection of AI-Generated Images (Supplemental Materials),"['ai_safety', 'computer_vision', 'generative_ai']",generative_ai,['generative_ai'],[],False
++2966462029,Tracking Community Consistency in Dynamic Networks: An Influence-based Approach,"['database_systems', 'generative_ai']",generative_ai,['database_systems'],['algorithms'],False
++7167727964,Navigating the Future of Dance Education with GenAI,['generative_ai'],generative_ai,['hci'],[],False
++7125473265,The moral in the machine: Rentier mechanisms in content recommender system on digital platforms,"['generative_ai', 'information_retrieval']",generative_ai,['ai_safety'],[],False
++7135177077,AI-enabled blended teaching in Information Theory and Coding: an outcomes-based mixed-methods approach,['generative_ai'],generative_ai,['hci'],[],False
++7215117915,RECONSTRUCTING AUTHENTICITY IN THE AGE OF GENERATIVE ARTIFICIAL INTELLIGENCE: A QUALITATIVE CONTENT ANALYSIS OF AI-ASSISTED INSTAGRAM POSTS,['generative_ai'],generative_ai,['hci'],[],False
++4395661446,A Two-Stage Strategy for Retrieving 2-D Ocean Wave Spectra From Chinese Gaofen-3 SAR Wave Mode Products,['generative_ai'],generative_ai,"['computer_vision', 'generative_ai']",['computer_vision'],False
++7171857204,The chain-mediating roles of generative AI literacy and TPACK in the relationship between AI support and teaching practice among university teachers,['generative_ai'],generative_ai,[],[],False
++7143415918,Applications of Generative Artificial Intelligence in Business Management,['generative_ai'],generative_ai,[],[],False
++7112438470,Exploring emergent generative AI (GenAI) tools and their influences on architecture and built environment education.,['generative_ai'],generative_ai,[],[],False
++4387666097,A local color transfer method based on optimal transmission,"['computer_vision', 'generative_ai']",generative_ai,['computer_graphics'],['computer_vision'],False
++7118194253,Improving AI Literacy among Adult Learners through Basic Generative Artificial Intelligence Education and Its Application to Beauty Design,['generative_ai'],generative_ai,['hci'],[],False
++4417167881,Exploring AI-powered Digital Innovations from A Transnational Governance Perspective: Implications for Market Acceptance and Digital Accountability Accountability,"['ai_safety', 'computer_networks', 'hci']",ai_safety,[],[],False
++2910154596,"Ethical Implications: The ACM/IEEE-CS Software Engineering Code applied to Tesla's ""Autopilot"" System","['ai_safety', 'software_engineering']",ai_safety,[],['software_engineering'],False
++7133246977,"Exploring Social Sciences and Humanities Student Perspectives on AI in Higher Education: A Comprehensive Study of Usage, Ethics, and Future Intentions","['ai_safety', 'privacy_preserving_computation']",ai_safety,[],[],False
++7119528090,Criminal Liability of Generative Artificial Intelligence Providers for User-Generated Child Sexual Abuse Material,"['ai_safety', 'generative_ai', 'software_engineering']",ai_safety,[],['ai_safety'],False
++7168242482,Ethical and Societal Impacts of Generative AI in Higher Computing Education: An ACM Task Force Working Group to Develop a Landscape Analysis - Perspectives from the Global Souths and Guidelines for CS1/CS2/CS3,"['ai_safety', 'database_systems', 'generative_ai']",ai_safety,[],[],False
++4399042091,Rethinking Health Recommender Systems for Active Aging: An Autonomy-Based Ethical Analysis,"['ai_safety', 'information_retrieval']",ai_safety,['hci'],['ai_safety'],False
++7214017973,From AI ethics principles to academic design practice: a faculty-led generative AI case in doctoral education,"['ai_safety', 'generative_ai']",ai_safety,[],[],False
++2803572453,An Exploration of Verbatim Content Republishing by News Producers,['ai_safety'],ai_safety,"['information_retrieval', 'nlp']",['information_retrieval'],False
++3118441209,Web Analytics and Online Retail,['ai_safety'],ai_safety,['hci'],['privacy_preserving_computation'],False
++7161719656,Are Researchers Being Replaced by Artificial Intelligence?,"['ai_agents', 'ai_safety']",ai_safety,['autonomous_scientific_discovery'],['autonomous_scientific_discovery'],False
++7164757242,Risk Governance of Algorithmic Recommendation in Education for Fostering a Strong Sense of Community for the Chinese Nation: A Conceptual Analysis in Chinese Higher Education,"['ai_safety', 'algorithms', 'information_retrieval']",ai_safety,[],['ai_safety'],False
++7119498334,"From Abstract Threats to Institutional Realities: A Comparative Semantic Network Analysis of AI Securitisation in the US, EU, and China",['ai_safety'],ai_safety,[],['ai_safety'],False
++7221224735,Optimizing NotebookLM for Islamic Religious Education in the Age of Artificial Intelligence,['ai_safety'],ai_safety,[],[],False
++4400082893,القضايا الأمنية والأخلاقية لاستخدام الذكاء الاصطناعي وعلاقتها بالاتجاه نحو تطبيقاته لدى طلاب الجامعة,['ai_safety'],ai_safety,[],[],False
++4394613885,تطوير مبادرة المجتمع المهني للذكاء الاصطناعي في وزارة التربية والتعليم بسلطنة عمان في ضوء نموذج الاتحاد الأوروبي,['ai_safety'],ai_safety,[],[],False
++4411990784,Investigation and Research on University Students’ Awareness of Artificial Intelligence,['ai_safety'],ai_safety,[],['hci'],False
++4220714491,Informatics and Society in German Computer Sciences Bachelor Courses,['ai_safety'],ai_safety,[],[],False
++3096804093,Recommendations for artificial intelligence implementation in African governments,['ai_safety'],ai_safety,[],['ai_safety'],True
++4405384079,Making academia suck less: Supporting early career researchers studying harmful content online through a feminist ethics of care,['ai_safety'],ai_safety,[],['hci'],False
++3093512442,Best practices for authors of healthcare-related artificial intelligence manuscripts,['ai_safety'],ai_safety,[],['ai_safety'],False
++4415829954,Artificial intelligence competency among academics in Nigeria: A roadmap for ethical usage in the fifth industrial revolution (5IR),"['ai_safety', 'llm']",ai_safety,[],[],True
++4392109933,The Future of Research in an Artificial Intelligence-Driven World,['ai_safety'],ai_safety,[],['autonomous_scientific_discovery'],False
++7211974235,"ARTIFICIAL INTELLIGENCE AND HUMAN COMMUNICATION IN EDUCATION AND SOCIETY: RECONFIGURING LANGUAGE, MEDIA, AND SOCIAL INTERACTION IN THE DIGITAL AGE","['ai_safety', 'algorithms', 'generative_ai', 'information_retrieval', 'nlp']",ai_safety,[],[],False
++4283010037,Minorities in networks and algorithms,"['ai_safety', 'information_retrieval']",ai_safety,[],['ai_safety'],False
++4412611638,Analogical reasoning as a core AGI capability,['ai_safety'],ai_safety,['ai_agents'],[],False
++4290803082,"Decolonisation, Global Data Law, and Indigenous Data Sovereignty","['ai_safety', 'privacy_preserving_computation']",ai_safety,[],[],False
++7204733568,Bridging Attitudes and AI Policy: Stakeholder Preferences for AI Integration in U.S. Hybrid K–12 Education,['ai_safety'],ai_safety,[],[],False
++4399331704,AI development in financial markets: a balanced scorecard analysis of its impact on sustainable development goals (February 2024),['ai_safety'],ai_safety,[],[],False
++7164020700,Yapay Zeka ve Estetikte ‘Kopyala-Yapıştır’ Tehlikesi,['ai_safety'],ai_safety,['generative_ai'],['generative_ai'],False
++2997003061,Incorporating Model Based Safety Design into Development Process of Civil Aircraft Systems,['ai_safety'],ai_safety,['ai_safety'],['software_engineering'],True
++7133303346,16.2 A Bias-Flip-Based Piezoelectric Energy Harvesting Interface with a Digital Track-and-Lock MPPT Achieving Sampling-Free Operation and 99.8% MPPT Efficiency,['distributed_computing'],distributed_computing,['computer_architecture'],['computer_architecture'],False
++2950925562,Localization results for Minkowski contents,['distributed_computing'],distributed_computing,[],[],False
++3115601909,Incentive-based Peer-to-Peer Distributed Energy Trading in Smart Grid Systems,"['blockchain', 'computer_networks', 'distributed_computing']",distributed_computing,['sustainable_computing'],['algorithms'],False
++4384570508,Enhancing Scalability through Extended Hybrid Switching with Mixture Path Deployment (EHS-MPD): A Novel Framework,"['cloud_computing', 'computer_networks', 'database_systems', 'distributed_computing', 'operating_systems']",distributed_computing,['database_systems'],"['computer_networks', 'distributed_computing']",False
++7128410034,Résolution de systèmes polynomiaux à paramètres par Représentation Univariée Rationnelle Générique,['distributed_computing'],distributed_computing,['algorithms'],['algorithms'],False
++2860211826,Fractional order control of the cyber-physical cryogenic isotope separation columns cascade system,['distributed_computing'],distributed_computing,[],[],False
++2907696829,Energy Management System in Micro-Grid with Storage and Hydrogen Production,"['computer_architecture', 'cybersecurity', 'distributed_computing']",distributed_computing,[],[],False
++2805585446,Exploring Academic Performance Paths and Student Learning Strategies in a Large Foundational Engineering Course,['distributed_computing'],distributed_computing,[],[],False
++4409660083,Model-Free Predictive Control Strategy for Inverter Based on Adaptive Super- Twisting Sliding Mode Observer,"['applied_cryptography', 'distributed_computing']",distributed_computing,[],[],False
++7161207400,SisMob renforce son parc instrumental avec 500 nouveaux nodes,"['computer_networks', 'distributed_computing']",distributed_computing,[],[],False
++2904849773,Constructing Business Simulation Training Platform Based on Workflow Management Systems,['distributed_computing'],distributed_computing,['hci'],['software_engineering'],False
++4401163816,Frequency Estimation Method for Increasing Accuracy of PM by Combining Multipoint Software Zero-Crossing Recognition and Least Square Regression,['distributed_computing'],distributed_computing,[],[],False
++4200525533,The establishment and evaluation of the automatic crisis balance analysis model for social network users based on artificial intelligence technology,['distributed_computing'],distributed_computing,"['ai_safety', 'nlp']",['nlp'],False
++3038439485,Development of Pop-Up Book Media based on Balanced Literacy Approach to Improve Skills of Reading Class 1 Students Basic School,['distributed_computing'],distributed_computing,[],[],False
++3014911997,Finite-time synchronization for fractional-order memristor-based neural networks with discontinuous activations and multiple delays,"['ai_accelerators', 'distributed_computing']",distributed_computing,['ml_theory'],[],False
++2903036202,Synchronization of discrete-time Lur’e systems under saturating control,"['distributed_computing', 'programming_languages']",distributed_computing,['ml_theory'],[],False
++4286377584,Empirical Bayesian Estimation of the Interferometric SAR Coherence Magnitude,['distributed_computing'],distributed_computing,"['ml_theory', 'scientific_computing']",['scientific_computing'],False
++7131086620,"Assessing Artificial Intelligence in Breast Imaging: A Survey of Breast Radiologists’ Insights on Adoption, Benefits, and Challenges",['distributed_computing'],distributed_computing,[],[],False
++3045355300,Distributed adaptive consensus protocol with decaying gains,"['ai_agents', 'computer_networks', 'distributed_computing', 'scientific_computing']",distributed_computing,['ml_theory'],[],False
++2903731864,A simulation study for count data models under varying degrees of outliers and zeros,['distributed_computing'],distributed_computing,[],[],False
++4200552233,Controlling the Runtime Overhead of Python Monitoring with Selective Instrumentation,"['computer_architecture', 'distributed_computing', 'scientific_computing']",distributed_computing,['software_engineering'],"['distributed_computing', 'software_engineering']",False
++7216249722,INTEGRASI NILAI PROFETIK DALAM E-MODUL INTERAKTIF MENULIS CERPEN UNTUK MENINGKATKAN LITERASI KREATIF SISWA,['distributed_computing'],distributed_computing,[],[],False
++4403092717,Parallel processing of geographical data in a mobile application,"['computer_architecture', 'distributed_computing']",distributed_computing,['distributed_computing'],['edge_computing'],False
++3123252765,Combining Input Shaping and Adaptive Model-Following Control for Vibration Suppression,['distributed_computing'],distributed_computing,['robotics'],['robotics'],False
++3093694272,The nonlinear influence of harmonious information technology affordance on organisational innovation,['distributed_computing'],distributed_computing,[],[],False
++3048509415,"Product Quality Improvement Policies in Industry 4.0: Characteristics, Enabling Factors, Barriers, and Evolution Toward Zero Defect Manufacturing","['distributed_computing', 'scientific_computing']",distributed_computing,[],[],False
++4392858396,Engaging Students from Rural Communities in AI Education with Game-Based Learning,"['nlp', 'reinforcement_learning']",reinforcement_learning,"['hci', 'nlp']",['hci'],False
++4405358450,Digital Open-Access and Open Educational Resources for the Music History Classroom,"['applied_cryptography', 'cybersecurity', 'distributed_computing', 'reinforcement_learning']",reinforcement_learning,[],[],False
++7160554706,A brief history of dopamine prediction errors,"['computational_biology', 'reinforcement_learning']",reinforcement_learning,[],"['computational_biology', 'reinforcement_learning']",False
++4393861500,A Reversible Investment Problem with Capacity and Demand in Finite Horizon: Free Boundary Analysis,"['reinforcement_learning', 'scientific_computing', 'sustainable_computing']",reinforcement_learning,[],[],False
++4220992898,Artificial Multisensory Neurons with Fused Haptic and Temperature Perception for Multimodal In‐Sensor Computing,"['ai_accelerators', 'hci', 'robotics']",ai_accelerators,"['computer_architecture', 'deep_learning_architectures', 'robotics']",['ai_accelerators'],False
++4416096548,AI-Driven “Spintronics Industry + Law” Curriculum Reconstruction,['ai_accelerators'],ai_accelerators,[],[],True
++4392976059,Brief Reviews of Books and Products: NASA TOPS Open Science 101,['ai_accelerators'],ai_accelerators,[],[],False
++3193775649,Stochastic Memristive Interface for Neural Signal Processing,['ai_accelerators'],ai_accelerators,['computer_architecture'],['computer_architecture'],False
++4408280919,Coupling Heterogeneous Neural Networks With Memristors,"['ai_accelerators', 'applied_cryptography', 'deep_learning_architectures', 'scientific_computing']",ai_accelerators,['computer_architecture'],"['applied_cryptography', 'cybersecurity']",False
++3039564587,Influence of the magnetic field on dielectric breakdown in memristors based on h-BN stacks,['ai_accelerators'],ai_accelerators,['computer_architecture'],[],False
++2955631546,Origin of Current‐Controlled Negative Differential Resistance Modes and the Emergence of Composite Characteristics with High Complexity,"['ai_accelerators', 'computer_architecture']",ai_accelerators,['computer_architecture'],[],False
++2846788133,Simple chaotic oscillators with diode bridges,['ai_accelerators'],ai_accelerators,['computer_architecture'],['computer_architecture'],False
++4391086735,On-Chip Universal Linear Optics using a 4x4 Silicon Photonic Coherent Crossbar,"['ai_accelerators', 'quantum_computing']",ai_accelerators,['computer_architecture'],"['ai_accelerators', 'computer_architecture']",False
++7171493874,Variational processing of partially coherent light in a photonic integrated circuit,['ai_accelerators'],ai_accelerators,['computer_architecture'],['ai_accelerators'],False
++4386996793,"Study of the Interseismic Deformation and Locking Depth along the Xidatan–Dongdatan Segment of the East Kunlun Fault Zone, Northeast Qinghai–Tibet Plateau, Based on Sentinel-1 Interferometry",['ai_accelerators'],ai_accelerators,[],[],True
++2801359411,Data Driven Models of Short-Term Synaptic Plasticity,['ai_accelerators'],ai_accelerators,['computational_biology'],['computational_biology'],False
++2946793717,Parasitic Memcapacitor Effects on HP TiO 2 Memristor Dynamics,['ai_accelerators'],ai_accelerators,['computer_architecture'],[],False
++3112050902,Quasisynchronization of Memristive Neural Networks With Communication Delays via Event-Triggered Impulsive Control,['ai_accelerators'],ai_accelerators,['ml_theory'],[],False
++4383046458,A Unified Framework for Online Data-Driven Predictive Control with Robust Safety Guarantees,"['applied_cryptography', 'computational_complexity']",computational_complexity,"['ai_safety', 'ml_theory', 'reinforcement_learning']","['ml_theory', 'robotics']",False
++2795888755,Generators and presentations for direct and wreath products of monoid acts,['computational_complexity'],computational_complexity,['algorithms'],[],False
++2979317633,Non-invasive assessments of the advanced glycation end products in human skin using reflectance NIR spectroscopy,['computational_complexity'],computational_complexity,[],[],False
++3037346090,Generalized Inverse-Free Kalman filter using Approximate Inverse of Block Diagonally Dominant Matrices,['computational_complexity'],computational_complexity,['algorithms'],['scientific_computing'],False
++4312278247,Measures for the Degree of Time Variance: Application to Filter Banks,"['computational_complexity', 'distributed_computing']",computational_complexity,['deep_learning_architectures'],[],False
++2928354599,Co-evolving cellular automata for morphogenesis,['computational_complexity'],computational_complexity,['computational_biology'],"['computational_biology', 'scientific_computing']",False
++2791221456,Prediksi Tingkat Kerawanan Kebakaran di Daerah Kudus Menggunakan Fuzzy Tsukamoto,['computational_complexity'],computational_complexity,['ml_theory'],[],False
++4410227679,Improved IMM Algorithm Based on Deep Learning for Maneuvering Target Tracking,['computational_complexity'],computational_complexity,"['ml_theory', 'robotics']","['deep_learning_architectures', 'robotics']",False
++4323338639,Efficient ADMM-Based Algorithm for Regularized Minimax Approximation,"['computational_complexity', 'computer_architecture']",computational_complexity,"['algorithms', 'deep_learning_architectures']",['algorithms'],False
++7134076622,"Optimization in offshore wind farms – a survey of problems, models, and solution approaches",['computational_complexity'],computational_complexity,['algorithms'],['algorithms'],False
++7125235431,Inferring relationships between petrophysical data and hyperspectral-derived mineralogy using PLS-based modelling,['computational_complexity'],computational_complexity,[],[],False
++7162015013,"The Influence of Visionary Leadership, Digital Marketing, and Digital Culture on Competitiveness at Institut Agama Islam Negeri (IAIN) Takengon",['computational_complexity'],computational_complexity,[],[],False
++4361029879,Distributed and Localized Model Predictive Control—Part II: Theoretical Guarantees,"['applied_cryptography', 'computational_complexity']",computational_complexity,['ml_theory'],['distributed_computing'],False
++4387624329,Effinformer: A Deep-Learning-Based Data-Driven Modeling of DC–DC Bidirectional Converters,['computational_complexity'],computational_complexity,['deep_learning_architectures'],['deep_learning_architectures'],False
++2794110548,Multi‐level thinking cellular automata using granular computing title,['computational_complexity'],computational_complexity,['ml_theory'],['algorithms'],True
++2990666756,Internal Model Control with Improved Butterworth Filter Based on Inverted Decoupling for Multivariable Systems,"['computational_complexity', 'distributed_computing']",computational_complexity,[],[],False
++2888657220,A Proportionate NLMS Algorithm for the Identification of Sparse Bilinear Forms,"['computational_complexity', 'cybersecurity']",computational_complexity,[],['algorithms'],False
++2968152157,Impact of airborne geophysical surveys on the exploration of gold mineralization in Burkina Faso,['computational_complexity'],computational_complexity,['scientific_computing'],[],True
++4414170577,Fully Actuated Approach for Safety-Critical Control of Underactuated Systems via Differential Flatness,"['computational_complexity', 'formal_verification']",computational_complexity,['robotics'],['robotics'],False
++2905468286,Laboratory Equipment and Supplies in the Environmental Microbiology Laboratory,['computational_complexity'],computational_complexity,[],[],False
++4396536343,Reduced-complexity multiple parameters estimation via toeplitz matrix triple iteration reconstruction with bistatic MIMO radar,['computational_complexity'],computational_complexity,['computer_architecture'],[],False
++4288109977,Modelling of corrosion using cellular automata,['computational_complexity'],computational_complexity,['scientific_computing'],['scientific_computing'],False
++2955693445,Absolute Root Separation,['computational_complexity'],computational_complexity,['computational_complexity'],['algorithms'],False
++4409570626,Quantum systems from random probabilistic automata,"['computational_complexity', 'scientific_computing']",computational_complexity,['quantum_computing'],['quantum_computing'],False
++7214073407,Finite Bases for Truncated Cyclic Group Flat Semirings with Two Independent Parameters,['computational_complexity'],computational_complexity,['formal_verification'],[],False
++2809671584,Random Ordering of Semiprimes,['computational_complexity'],computational_complexity,[],[],False
++4206034737,Artificial intelligence and remote sensing for spatial prediction of daily air temperature: case study of Souss watershed of Morocco,['computational_complexity'],computational_complexity,['computer_vision'],[],False
++7169788258,Time–Frequency Image-Based Overlapping LPI Radar Signal Detection and Recognition with LPI-YOLO,['computational_complexity'],computational_complexity,"['computer_vision', 'speech_audio']","['computer_vision', 'deep_learning_architectures']",False
++4409702353,Adoption of Massive Open Online Courses (MOOCs) for Health Informatics and Administration Sustainability Education in Saudi Arabia,"['computational_biology', 'computational_complexity']",computational_complexity,[],[],False
++3086506771,Ordered semihypergroup constructions,['computational_complexity'],computational_complexity,[],[],False
++7154114167,ANALISIS DESAIN FUZZY MAMDANI UNTUK KLASIFIKASI KUALITAS AIR BIOFLOK BERBASIS IOT,['computational_complexity'],computational_complexity,['hci'],['edge_computing'],False
++4313639087,Tracking Control of Multi-Agent Systems Using a Networked Predictive PID Tracking Scheme,['ai_agents'],ai_agents,"['ai_agents', 'ml_theory']",['distributed_computing'],False
++7171217691,Social Sustainability in the Agent-to-Agent economy: Artificial Intelligence and the Future of Tourism and Hospitality Labour,"['ai_agents', 'hci']",ai_agents,[],[],False
++4281774473,Extração e seleção de características para a classificação eficiente de séries temporais,['ai_agents'],ai_agents,['database_systems'],[],False
++2968942120,Transparência para Humanos e Máquinas: Um framework para Publicar Dados Abertos Interconectados Semanticamente Descritos,['ai_agents'],ai_agents,['database_systems'],['database_systems'],False
++7130252400,Adoção da Inteligência Artificial no Marketing Digital: um estudo sobre a percepção de gestores de pequenos negócios em Portugal,['ai_agents'],ai_agents,[],[],True
++2911791709,Desarrollo de acciones formativas a través del empleo de tecnologías e-learning. Aplicación a un curso de administración de sistemas de almacenamiento Netapp,['ai_agents'],ai_agents,[],[],False
++4412987259,Extended State Observer-Based Data-Driven Control of Multi-Agent Systems,"['ai_agents', 'robotics']",ai_agents,['ml_theory'],[],True
++3128896073,THEORY THROUGH ARGUMENT: APPLYING ARGUMENT MAPPING TO FACILITATE THEORY BUILDING,['ai_agents'],ai_agents,[],[],False
++3037034904,Impact of COVID-19 behavioral inertia on reopening strategies for New York City transit,['ai_agents'],ai_agents,[],[],False
++3014047513,A New Perspective to Algebraic Characterization on Controllability of Multiagent Systems,['ai_agents'],ai_agents,['ml_theory'],[],False
++3122776204,Containment Control for Upper-Triangular Nonlinear Multi-Agent Systems Perturbed by Second-Order Moment Processes,['ai_agents'],ai_agents,[],[],False
++7160854959,Dynamic event-triggered prescribed performance control for nonlinear multi-agent systems: a finite-time convergence strategy,['ai_agents'],ai_agents,['ml_theory'],[],False
++4212806673,Influence of Flow and Spatial Heterogeneities on Pattern Formation of Dictyostelium discoideum,"['ai_agents', 'scientific_computing']",ai_agents,['computational_biology'],[],False
++4402920523,Distributed Bipartite State Consensus Tracking for Discrete-Time Singular Multi-agent Systems,['ai_agents'],ai_agents,['ml_theory'],[],True
++7137881325,Inteligência artificial no ensino de redação do ENEM,['ai_agents'],ai_agents,"['llm', 'nlp']","['hci', 'llm']",False
++7218192222,Χρήση απλών μορφών πολυ-πρακτορικών συστημάτων για τη διδασκαλία βασικών ιδιοτήτων των πολύπλοκων συστημάτων,['ai_agents'],ai_agents,[],['ai_agents'],False
++2898995053,Inherent Game Characteristics of Electronic Negotiations,['ai_agents'],ai_agents,['hci'],['hci'],False
++4362606359,Reasoning about causality in games,['ai_agents'],ai_agents,['ml_theory'],"['ai_agents', 'ml_theory']",False
++4390619203,TRANSFORMAÇÃO EDUCACIONAL: O ASCENDENTE ENSINO A DISTÂNCIA NO BRASIL,['ai_agents'],ai_agents,[],[],False
++2898910837,Adaptive Fuzzy Consensus Control for Unknown State-Delay Nonlinear Multiagent System,['ai_agents'],ai_agents,"['ai_safety', 'reinforcement_learning']",['ai_agents'],False
++7219983900,Do húmus ao humano: o regime de informação fundiário brasileiro,['ai_agents'],ai_agents,[],[],False
++7132739348,BKLM - An expressive logic for defeasible reasoning,['ai_agents'],ai_agents,['ml_theory'],['formal_verification'],False
++2920799053,Rise of bots,"['ai_agents', 'distributed_computing']",ai_agents,[],['algorithms'],True
++2993634112,Mobile learning e pensamento computacional: contributos para o desenvolvimento de aplicações em contextos educativos,"['ai_agents', 'software_engineering']",ai_agents,['hci'],['hci'],True
++7214219730,The finite-time bounded consensus of general linear discrete-time multi-agent systems,['ai_agents'],ai_agents,['ml_theory'],[],False

added taxonomy/validation/metrics.json +997 −0

++{
++ "n_rated_by_all": 2181,
++ "raters": [
++  "claude-haiku-5-5",
++  "claude-sonnet-5-5"
++ ],
++ "rules_vs_claude-haiku-5-5": {
++  "per_category": {
++   "llm": {
++    "tp": 95,
++    "fp": 86,
++    "fn": 10,
++    "precision": 0.5248618784530387,
++    "recall": 0.9047619047619048,
++    "f1": 0.6643356643356644
++   },
++   "foundation_models": {
++    "tp": 23,
++    "fp": 108,
++    "fn": 1,
++    "precision": 0.17557251908396945,
++    "recall": 0.9583333333333334,
++    "f1": 0.2967741935483871
++   },
++   "multimodal_ai": {
++    "tp": 56,
++    "fp": 30,
++    "fn": 15,
++    "precision": 0.6511627906976745,
++    "recall": 0.7887323943661971,
++    "f1": 0.7133757961783439
++   },
++   "generative_ai": {
++    "tp": 57,
++    "fp": 53,
++    "fn": 17,
++    "precision": 0.5181818181818182,
++    "recall": 0.7702702702702703,
++    "f1": 0.6195652173913043
++   },
++   "ai_agents": {
++    "tp": 42,
++    "fp": 51,
++    "fn": 21,
++    "precision": 0.45161290322580644,
++    "recall": 0.6666666666666666,
++    "f1": 0.5384615384615384
++   },
++   "reinforcement_learning": {
++    "tp": 78,
++    "fp": 25,
++    "fn": 15,
++    "precision": 0.7572815533980582,
++    "recall": 0.8387096774193549,
++    "f1": 0.7959183673469389
++   },
++   "ml_theory": {
++    "tp": 81,
++    "fp": 122,
++    "fn": 110,
++    "precision": 0.39901477832512317,
++    "recall": 0.42408376963350786,
++    "f1": 0.4111675126903553
++   },
++   "deep_learning_architectures": {
++    "tp": 111,
++    "fp": 194,
++    "fn": 37,
++    "precision": 0.3639344262295082,
++    "recall": 0.75,
++    "f1": 0.49006622516556303
++   },
++   "computer_vision": {
++    "tp": 94,
++    "fp": 161,
++    "fn": 47,
++    "precision": 0.3686274509803922,
++    "recall": 0.6666666666666666,
++    "f1": 0.4747474747474748
++   },
++   "nlp": {
++    "tp": 88,
++    "fp": 116,
++    "fn": 27,
++    "precision": 0.43137254901960786,
++    "recall": 0.7652173913043478,
++    "f1": 0.5517241379310345
++   },
++   "speech_audio": {
++    "tp": 49,
++    "fp": 24,
++    "fn": 12,
++    "precision": 0.6712328767123288,
++    "recall": 0.8032786885245902,
++    "f1": 0.7313432835820896
++   },
++   "robotics": {
++    "tp": 64,
++    "fp": 36,
++    "fn": 19,
++    "precision": 0.64,
++    "recall": 0.7710843373493976,
++    "f1": 0.6994535519125684
++   },
++   "hci": {
++    "tp": 59,
++    "fp": 77,
++    "fn": 96,
++    "precision": 0.4338235294117647,
++    "recall": 0.38064516129032255,
++    "f1": 0.4054982817869416
++   },
++   "software_engineering": {
++    "tp": 55,
++    "fp": 128,
++    "fn": 17,
++    "precision": 0.3005464480874317,
++    "recall": 0.7638888888888888,
++    "f1": 0.4313725490196078
++   },
++   "programming_languages": {
++    "tp": 34,
++    "fp": 79,
++    "fn": 2,
++    "precision": 0.3008849557522124,
++    "recall": 0.9444444444444444,
++    "f1": 0.45637583892617456
++   },
++   "compilers": {
++    "tp": 26,
++    "fp": 46,
++    "fn": 1,
++    "precision": 0.3611111111111111,
++    "recall": 0.9629629629629629,
++    "f1": 0.5252525252525253
++   },
++   "formal_verification": {
++    "tp": 41,
++    "fp": 47,
++    "fn": 3,
++    "precision": 0.4659090909090909,
++    "recall": 0.9318181818181818,
++    "f1": 0.6212121212121211
++   },
++   "algorithms": {
++    "tp": 46,
++    "fp": 132,
++    "fn": 26,
++    "precision": 0.25842696629213485,
++    "recall": 0.6388888888888888,
++    "f1": 0.368
++   },
++   "computational_complexity": {
++    "tp": 25,
++    "fp": 82,
++    "fn": 13,
++    "precision": 0.2336448598130841,
++    "recall": 0.6578947368421053,
++    "f1": 0.3448275862068965
++   },
++   "database_systems": {
++    "tp": 56,
++    "fp": 93,
++    "fn": 27,
++    "precision": 0.37583892617449666,
++    "recall": 0.6746987951807228,
++    "f1": 0.48275862068965525
++   },
++   "information_retrieval": {
++    "tp": 54,
++    "fp": 59,
++    "fn": 30,
++    "precision": 0.4778761061946903,
++    "recall": 0.6428571428571429,
++    "f1": 0.5482233502538072
++   },
++   "distributed_computing": {
++    "tp": 58,
++    "fp": 208,
++    "fn": 13,
++    "precision": 0.21804511278195488,
++    "recall": 0.8169014084507042,
++    "f1": 0.344213649851632
++   },
++   "cloud_computing": {
++    "tp": 62,
++    "fp": 88,
++    "fn": 4,
++    "precision": 0.41333333333333333,
++    "recall": 0.9393939393939394,
++    "f1": 0.5740740740740741
++   },
++   "edge_computing": {
++    "tp": 44,
++    "fp": 65,
++    "fn": 6,
++    "precision": 0.4036697247706422,
++    "recall": 0.88,
++    "f1": 0.5534591194968553
++   },
++   "computer_networks": {
++    "tp": 57,
++    "fp": 125,
++    "fn": 10,
++    "precision": 0.3131868131868132,
++    "recall": 0.8507462686567164,
++    "f1": 0.4578313253012048
++   },
++   "cybersecurity": {
++    "tp": 102,
++    "fp": 141,
++    "fn": 20,
++    "precision": 0.41975308641975306,
++    "recall": 0.8360655737704918,
++    "f1": 0.5589041095890411
++   },
++   "applied_cryptography": {
++    "tp": 54,
++    "fp": 82,
++    "fn": 0,
++    "precision": 0.39705882352941174,
++    "recall": 1.0,
++    "f1": 0.5684210526315789
++   },
++   "blockchain": {
++    "tp": 53,
++    "fp": 32,
++    "fn": 0,
++    "precision": 0.6235294117647059,
++    "recall": 1.0,
++    "f1": 0.7681159420289855
++   },
++   "computer_architecture": {
++    "tp": 57,
++    "fp": 162,
++    "fn": 54,
++    "precision": 0.2602739726027397,
++    "recall": 0.5135135135135135,
++    "f1": 0.3454545454545454
++   },
++   "ai_accelerators": {
++    "tp": 36,
++    "fp": 66,
++    "fn": 5,
++    "precision": 0.35294117647058826,
++    "recall": 0.8780487804878049,
++    "f1": 0.5034965034965035
++   },
++   "operating_systems": {
++    "tp": 19,
++    "fp": 84,
++    "fn": 1,
++    "precision": 0.18446601941747573,
++    "recall": 0.95,
++    "f1": 0.3089430894308943
++   },
++   "quantum_computing": {
++    "tp": 48,
++    "fp": 28,
++    "fn": 3,
++    "precision": 0.631578947368421,
++    "recall": 0.9411764705882353,
++    "f1": 0.7559055118110236
++   },
++   "scientific_computing": {
++    "tp": 56,
++    "fp": 117,
++    "fn": 34,
++    "precision": 0.3236994219653179,
++    "recall": 0.6222222222222222,
++    "f1": 0.42585551330798477
++   },
++   "computational_biology": {
++    "tp": 50,
++    "fp": 58,
++    "fn": 44,
++    "precision": 0.46296296296296297,
++    "recall": 0.5319148936170213,
++    "f1": 0.495049504950495
++   },
++   "computer_graphics": {
++    "tp": 34,
++    "fp": 46,
++    "fn": 8,
++    "precision": 0.425,
++    "recall": 0.8095238095238095,
++    "f1": 0.5573770491803278
++   },
++   "privacy_preserving_computation": {
++    "tp": 41,
++    "fp": 108,
++    "fn": 0,
++    "precision": 0.2751677852348993,
++    "recall": 1.0,
++    "f1": 0.43157894736842106
++   },
++   "federated_learning": {
++    "tp": 54,
++    "fp": 21,
++    "fn": 2,
++    "precision": 0.72,
++    "recall": 0.9642857142857143,
++    "f1": 0.8244274809160306
++   },
++   "ai_safety": {
++    "tp": 47,
++    "fp": 70,
++    "fn": 22,
++    "precision": 0.4017094017094017,
++    "recall": 0.6811594202898551,
++    "f1": 0.5053763440860215
++   },
++   "sustainable_computing": {
++    "tp": 28,
++    "fp": 56,
++    "fn": 4,
++    "precision": 0.3333333333333333,
++    "recall": 0.875,
++    "f1": 0.48275862068965514
++   },
++   "autonomous_scientific_discovery": {
++    "tp": 7,
++    "fp": 56,
++    "fn": 3,
++    "precision": 0.1111111111111111,
++    "recall": 0.7,
++    "f1": 0.19178082191780818
++   }
++  },
++  "micro_precision": 0.38906051244775575,
++  "micro_recall": 0.7332191780821918,
++  "micro_f1": 0.5083699394515018,
++  "macro_f1": 0.5205869260555518
++ },
++ "kappa_primary_rules_vs_claude-haiku-5-5": 0.45309445479816046,
++ "rules_vs_claude-sonnet-5-5": {
++  "per_category": {
++   "llm": {
++    "tp": 84,
++    "fp": 97,
++    "fn": 5,
++    "precision": 0.46408839779005523,
++    "recall": 0.9438202247191011,
++    "f1": 0.6222222222222221
++   },
++   "foundation_models": {
++    "tp": 27,
++    "fp": 104,
++    "fn": 0,
++    "precision": 0.20610687022900764,
++    "recall": 1.0,
++    "f1": 0.34177215189873417
++   },
++   "multimodal_ai": {
++    "tp": 59,
++    "fp": 27,
++    "fn": 14,
++    "precision": 0.686046511627907,
++    "recall": 0.8082191780821918,
++    "f1": 0.7421383647798743
++   },
++   "generative_ai": {
++    "tp": 55,
++    "fp": 55,
++    "fn": 15,
++    "precision": 0.5,
++    "recall": 0.7857142857142857,
++    "f1": 0.6111111111111112
++   },
++   "ai_agents": {
++    "tp": 42,
++    "fp": 51,
++    "fn": 15,
++    "precision": 0.45161290322580644,
++    "recall": 0.7368421052631579,
++    "f1": 0.56
++   },
++   "reinforcement_learning": {
++    "tp": 80,
++    "fp": 23,
++    "fn": 3,
++    "precision": 0.7766990291262136,
++    "recall": 0.963855421686747,
++    "f1": 0.8602150537634409
++   },
++   "ml_theory": {
++    "tp": 54,
++    "fp": 149,
++    "fn": 37,
++    "precision": 0.2660098522167488,
++    "recall": 0.5934065934065934,
++    "f1": 0.36734693877551017
++   },
++   "deep_learning_architectures": {
++    "tp": 92,
++    "fp": 213,
++    "fn": 28,
++    "precision": 0.3016393442622951,
++    "recall": 0.7666666666666667,
++    "f1": 0.4329411764705882
++   },
++   "computer_vision": {
++    "tp": 84,
++    "fp": 171,
++    "fn": 26,
++    "precision": 0.32941176470588235,
++    "recall": 0.7636363636363637,
++    "f1": 0.4602739726027398
++   },
++   "nlp": {
++    "tp": 82,
++    "fp": 122,
++    "fn": 6,
++    "precision": 0.4019607843137255,
++    "recall": 0.9318181818181818,
++    "f1": 0.5616438356164384
++   },
++   "speech_audio": {
++    "tp": 49,
++    "fp": 24,
++    "fn": 4,
++    "precision": 0.6712328767123288,
++    "recall": 0.9245283018867925,
++    "f1": 0.7777777777777778
++   },
++   "robotics": {
++    "tp": 61,
++    "fp": 39,
++    "fn": 16,
++    "precision": 0.61,
++    "recall": 0.7922077922077922,
++    "f1": 0.6892655367231638
++   },
++   "hci": {
++    "tp": 63,
++    "fp": 73,
++    "fn": 67,
++    "precision": 0.4632352941176471,
++    "recall": 0.4846153846153846,
++    "f1": 0.4736842105263158
++   },
++   "software_engineering": {
++    "tp": 68,
++    "fp": 115,
++    "fn": 24,
++    "precision": 0.37158469945355194,
++    "recall": 0.7391304347826086,
++    "f1": 0.49454545454545457
++   },
++   "programming_languages": {
++    "tp": 40,
++    "fp": 73,
++    "fn": 0,
++    "precision": 0.35398230088495575,
++    "recall": 1.0,
++    "f1": 0.522875816993464
++   },
++   "compilers": {
++    "tp": 32,
++    "fp": 40,
++    "fn": 0,
++    "precision": 0.4444444444444444,
++    "recall": 1.0,
++    "f1": 0.6153846153846153
++   },
++   "formal_verification": {
++    "tp": 51,
++    "fp": 37,
++    "fn": 4,
++    "precision": 0.5795454545454546,
++    "recall": 0.9272727272727272,
++    "f1": 0.7132867132867132
++   },
++   "algorithms": {
++    "tp": 59,
++    "fp": 119,
++    "fn": 54,
++    "precision": 0.33146067415730335,
++    "recall": 0.5221238938053098,
++    "f1": 0.40549828178694164
++   },
++   "computational_complexity": {
++    "tp": 24,
++    "fp": 83,
++    "fn": 3,
++    "precision": 0.22429906542056074,
++    "recall": 0.8888888888888888,
++    "f1": 0.3582089552238806
++   },
++   "database_systems": {
++    "tp": 43,
++    "fp": 106,
++    "fn": 11,
++    "precision": 0.28859060402684567,
++    "recall": 0.7962962962962963,
++    "f1": 0.4236453201970444
++   },
++   "information_retrieval": {
++    "tp": 52,
++    "fp": 61,
++    "fn": 3,
++    "precision": 0.46017699115044247,
++    "recall": 0.9454545454545454,
++    "f1": 0.6190476190476191
++   },
++   "distributed_computing": {
++    "tp": 43,
++    "fp": 223,
++    "fn": 12,
++    "precision": 0.16165413533834586,
++    "recall": 0.7818181818181819,
++    "f1": 0.26791277258566976
++   },
++   "cloud_computing": {
++    "tp": 70,
++    "fp": 80,
++    "fn": 1,
++    "precision": 0.4666666666666667,
++    "recall": 0.9859154929577465,
++    "f1": 0.6334841628959276
++   },
++   "edge_computing": {
++    "tp": 48,
++    "fp": 61,
++    "fn": 9,
++    "precision": 0.44036697247706424,
++    "recall": 0.8421052631578947,
++    "f1": 0.5783132530120482
++   },
++   "computer_networks": {
++    "tp": 61,
++    "fp": 121,
++    "fn": 4,
++    "precision": 0.33516483516483514,
++    "recall": 0.9384615384615385,
++    "f1": 0.49392712550607293
++   },
++   "cybersecurity": {
++    "tp": 105,
++    "fp": 138,
++    "fn": 11,
++    "precision": 0.43209876543209874,
++    "recall": 0.9051724137931034,
++    "f1": 0.5849582172701949
++   },
++   "applied_cryptography": {
++    "tp": 49,
++    "fp": 87,
++    "fn": 0,
++    "precision": 0.3602941176470588,
++    "recall": 1.0,
++    "f1": 0.5297297297297298
++   },
++   "blockchain": {
++    "tp": 57,
++    "fp": 28,
++    "fn": 0,
++    "precision": 0.6705882352941176,
++    "recall": 1.0,
++    "f1": 0.8028169014084507
++   },
++   "computer_architecture": {
++    "tp": 50,
++    "fp": 169,
++    "fn": 18,
++    "precision": 0.228310502283105,
++    "recall": 0.7352941176470589,
++    "f1": 0.3484320557491289
++   },
++   "ai_accelerators": {
++    "tp": 50,
++    "fp": 52,
++    "fn": 11,
++    "precision": 0.49019607843137253,
++    "recall": 0.819672131147541,
++    "f1": 0.6134969325153374
++   },
++   "operating_systems": {
++    "tp": 23,
++    "fp": 80,
++    "fn": 0,
++    "precision": 0.22330097087378642,
++    "recall": 1.0,
++    "f1": 0.3650793650793651
++   },
++   "quantum_computing": {
++    "tp": 52,
++    "fp": 24,
++    "fn": 3,
++    "precision": 0.6842105263157895,
++    "recall": 0.9454545454545454,
++    "f1": 0.7938931297709924
++   },
++   "scientific_computing": {
++    "tp": 48,
++    "fp": 125,
++    "fn": 17,
++    "precision": 0.2774566473988439,
++    "recall": 0.7384615384615385,
++    "f1": 0.40336134453781514
++   },
++   "computational_biology": {
++    "tp": 54,
++    "fp": 54,
++    "fn": 25,
++    "precision": 0.5,
++    "recall": 0.6835443037974683,
++    "f1": 0.5775401069518716
++   },
++   "computer_graphics": {
++    "tp": 38,
++    "fp": 42,
++    "fn": 2,
++    "precision": 0.475,
++    "recall": 0.95,
++    "f1": 0.6333333333333334
++   },
++   "privacy_preserving_computation": {
++    "tp": 56,
++    "fp": 93,
++    "fn": 5,
++    "precision": 0.37583892617449666,
++    "recall": 0.9180327868852459,
++    "f1": 0.5333333333333333
++   },
++   "federated_learning": {
++    "tp": 54,
++    "fp": 21,
++    "fn": 2,
++    "precision": 0.72,
++    "recall": 0.9642857142857143,
++    "f1": 0.8244274809160306
++   },
++   "ai_safety": {
++    "tp": 57,
++    "fp": 60,
++    "fn": 14,
++    "precision": 0.48717948717948717,
++    "recall": 0.8028169014084507,
++    "f1": 0.6063829787234042
++   },
++   "sustainable_computing": {
++    "tp": 32,
++    "fp": 52,
++    "fn": 1,
++    "precision": 0.38095238095238093,
++    "recall": 0.9696969696969697,
++    "f1": 0.547008547008547
++   },
++   "autonomous_scientific_discovery": {
++    "tp": 17,
++    "fp": 46,
++    "fn": 4,
++    "precision": 0.2698412698412698,
++    "recall": 0.8095238095238095,
++    "f1": 0.40476190476190477
++   }
++  },
++  "micro_precision": 0.39342176994366707,
++  "micro_recall": 0.8203865100416825,
++  "micro_f1": 0.531810366003439,
++  "macro_f1": 0.5548769450955702
++ },
++ "kappa_primary_rules_vs_claude-sonnet-5-5": 0.48350855645991686,
++ "kappa_primary_claude-haiku-5-5_vs_claude-sonnet-5-5": 0.7703253227683458,
++ "rules_vs_consensus": {
++  "per_category": {
++   "llm": {
++    "tp": 80,
++    "fp": 101,
++    "fn": 4,
++    "precision": 0.4419889502762431,
++    "recall": 0.9523809523809523,
++    "f1": 0.6037735849056604
++   },
++   "foundation_models": {
++    "tp": 20,
++    "fp": 111,
++    "fn": 0,
++    "precision": 0.15267175572519084,
++    "recall": 1.0,
++    "f1": 0.26490066225165565
++   },
++   "multimodal_ai": {
++    "tp": 55,
++    "fp": 31,
++    "fn": 9,
++    "precision": 0.6395348837209303,
++    "recall": 0.859375,
++    "f1": 0.7333333333333334
++   },
++   "generative_ai": {
++    "tp": 52,
++    "fp": 58,
++    "fn": 13,
++    "precision": 0.4727272727272727,
++    "recall": 0.8,
++    "f1": 0.5942857142857143
++   },
++   "ai_agents": {
++    "tp": 34,
++    "fp": 59,
++    "fn": 7,
++    "precision": 0.3655913978494624,
++    "recall": 0.8292682926829268,
++    "f1": 0.5074626865671642
++   },
++   "reinforcement_learning": {
++    "tp": 73,
++    "fp": 30,
++    "fn": 2,
++    "precision": 0.7087378640776699,
++    "recall": 0.9733333333333334,
++    "f1": 0.8202247191011237
++   },
++   "ml_theory": {
++    "tp": 48,
++    "fp": 155,
++    "fn": 26,
++    "precision": 0.23645320197044334,
++    "recall": 0.6486486486486487,
++    "f1": 0.34657039711191334
++   },
++   "deep_learning_architectures": {
++    "tp": 65,
++    "fp": 240,
++    "fn": 13,
++    "precision": 0.21311475409836064,
++    "recall": 0.8333333333333334,
++    "f1": 0.33942558746736295
++   },
++   "computer_vision": {
++    "tp": 74,
++    "fp": 181,
++    "fn": 22,
++    "precision": 0.2901960784313726,
++    "recall": 0.7708333333333334,
++    "f1": 0.4216524216524217
++   },
++   "nlp": {
++    "tp": 73,
++    "fp": 131,
++    "fn": 4,
++    "precision": 0.35784313725490197,
++    "recall": 0.948051948051948,
++    "f1": 0.5195729537366548
++   },
++   "speech_audio": {
++    "tp": 47,
++    "fp": 26,
++    "fn": 3,
++    "precision": 0.6438356164383562,
++    "recall": 0.94,
++    "f1": 0.7642276422764229
++   },
++   "robotics": {
++    "tp": 57,
++    "fp": 43,
++    "fn": 10,
++    "precision": 0.57,
++    "recall": 0.8507462686567164,
++    "f1": 0.6826347305389222
++   },
++   "hci": {
++    "tp": 51,
++    "fp": 85,
++    "fn": 51,
++    "precision": 0.375,
++    "recall": 0.5,
++    "f1": 0.42857142857142855
++   },
++   "software_engineering": {
++    "tp": 51,
++    "fp": 132,
++    "fn": 11,
++    "precision": 0.2786885245901639,
++    "recall": 0.8225806451612904,
++    "f1": 0.4163265306122449
++   },
++   "programming_languages": {
++    "tp": 30,
++    "fp": 83,
++    "fn": 0,
++    "precision": 0.26548672566371684,
++    "recall": 1.0,
++    "f1": 0.41958041958041964
++   },
++   "compilers": {
++    "tp": 26,
++    "fp": 46,
++    "fn": 0,
++    "precision": 0.3611111111111111,
++    "recall": 1.0,
++    "f1": 0.5306122448979591
++   },
++   "formal_verification": {
++    "tp": 36,
++    "fp": 52,
++    "fn": 1,
++    "precision": 0.4090909090909091,
++    "recall": 0.972972972972973,
++    "f1": 0.5760000000000001
++   },
++   "algorithms": {
++    "tp": 44,
++    "fp": 134,
++    "fn": 13,
++    "precision": 0.24719101123595505,
++    "recall": 0.7719298245614035,
++    "f1": 0.37446808510638296
++   },
++   "computational_complexity": {
++    "tp": 23,
++    "fp": 84,
++    "fn": 2,
++    "precision": 0.21495327102803738,
++    "recall": 0.92,
++    "f1": 0.3484848484848485
++   },
++   "database_systems": {
++    "tp": 39,
++    "fp": 110,
++    "fn": 7,
++    "precision": 0.26174496644295303,
++    "recall": 0.8478260869565217,
++    "f1": 0.4
++   },
++   "information_retrieval": {
++    "tp": 46,
++    "fp": 67,
++    "fn": 1,
++    "precision": 0.40707964601769914,
++    "recall": 0.9787234042553191,
++    "f1": 0.575
++   },
++   "distributed_computing": {
++    "tp": 34,
++    "fp": 232,
++    "fn": 4,
++    "precision": 0.12781954887218044,
++    "recall": 0.8947368421052632,
++    "f1": 0.2236842105263158
++   },
++   "cloud_computing": {
++    "tp": 56,
++    "fp": 94,
++    "fn": 1,
++    "precision": 0.37333333333333335,
++    "recall": 0.9824561403508771,
++    "f1": 0.5410628019323672
++   },
++   "edge_computing": {
++    "tp": 37,
++    "fp": 72,
++    "fn": 2,
++    "precision": 0.3394495412844037,
++    "recall": 0.9487179487179487,
++    "f1": 0.5
++   },
++   "computer_networks": {
++    "tp": 51,
++    "fp": 131,
++    "fn": 3,
++    "precision": 0.2802197802197802,
++    "recall": 0.9444444444444444,
++    "f1": 0.43220338983050843
++   },
++   "cybersecurity": {
++    "tp": 89,
++    "fp": 154,
++    "fn": 7,
++    "precision": 0.3662551440329218,
++    "recall": 0.9270833333333334,
++    "f1": 0.5250737463126843
++   },
++   "applied_cryptography": {
++    "tp": 42,
++    "fp": 94,
++    "fn": 0,
++    "precision": 0.3088235294117647,
++    "recall": 1.0,
++    "f1": 0.4719101123595506
++   },
++   "blockchain": {
++    "tp": 48,
++    "fp": 37,
++    "fn": 0,
++    "precision": 0.5647058823529412,
++    "recall": 1.0,
++    "f1": 0.7218045112781954
++   },
++   "computer_architecture": {
++    "tp": 45,
++    "fp": 174,
++    "fn": 15,
++    "precision": 0.2054794520547945,
++    "recall": 0.75,
++    "f1": 0.32258064516129037
++   },
++   "ai_accelerators": {
++    "tp": 35,
++    "fp": 67,
++    "fn": 2,
++    "precision": 0.3431372549019608,
++    "recall": 0.9459459459459459,
++    "f1": 0.5035971223021583
++   },
++   "operating_systems": {
++    "tp": 15,
++    "fp": 88,
++    "fn": 0,
++    "precision": 0.14563106796116504,
++    "recall": 1.0,
++    "f1": 0.2542372881355932
++   },
++   "quantum_computing": {
++    "tp": 45,
++    "fp": 31,
++    "fn": 3,
++    "precision": 0.5921052631578947,
++    "recall": 0.9375,
++    "f1": 0.7258064516129032
++   },
++   "scientific_computing": {
++    "tp": 38,
++    "fp": 135,
++    "fn": 8,
++    "precision": 0.21965317919075145,
++    "recall": 0.8260869565217391,
++    "f1": 0.34703196347031967
++   },
++   "computational_biology": {
++    "tp": 44,
++    "fp": 64,
++    "fn": 17,
++    "precision": 0.4074074074074074,
++    "recall": 0.7213114754098361,
++    "f1": 0.5207100591715976
++   },
++   "computer_graphics": {
++    "tp": 31,
++    "fp": 49,
++    "fn": 2,
++    "precision": 0.3875,
++    "recall": 0.9393939393939394,
++    "f1": 0.5486725663716815
++   },
++   "privacy_preserving_computation": {
++    "tp": 39,
++    "fp": 110,
++    "fn": 0,
++    "precision": 0.26174496644295303,
++    "recall": 1.0,
++    "f1": 0.41489361702127664
++   },
++   "federated_learning": {
++    "tp": 54,
++    "fp": 21,
++    "fn": 1,
++    "precision": 0.72,
++    "recall": 0.9818181818181818,
++    "f1": 0.8307692307692308
++   },
++   "ai_safety": {
++    "tp": 43,
++    "fp": 74,
++    "fn": 4,
++    "precision": 0.36752136752136755,
++    "recall": 0.9148936170212766,
++    "f1": 0.524390243902439
++   },
++   "sustainable_computing": {
++    "tp": 25,
++    "fp": 59,
++    "fn": 0,
++    "precision": 0.2976190476190476,
++    "recall": 1.0,
++    "f1": 0.4587155963302752
++   },
++   "autonomous_scientific_discovery": {
++    "tp": 7,
++    "fp": 56,
++    "fn": 3,
++    "precision": 0.1111111111111111,
++    "recall": 0.7,
++    "f1": 0.19178082191780818
++   }
++  },
++  "micro_precision": 0.32745775031800833,
++  "micro_recall": 0.8692715870718765,
++  "micro_f1": 0.4757127771911298,
++  "macro_f1": 0.49315080922219573
++ },
++ "krippendorff_alpha_masi_claude-haiku-5-5_vs_claude-sonnet-5-5": 0.6711402303757961,
++ "mean_jaccard_claude-haiku-5-5_vs_claude-sonnet-5-5": 0.5519983188139996,
++ "n_disagreements": 658
++}
+ \ No newline at end of file

modified taxonomy/validation/ratings_claude-haiku-5-5.jsonl +2191 −0

++{"work_id": 4400648397, "categories": ["computer_algebra_placeholder"], "primary": null, "confidence": 0.0, "rationale": "placeholder", "model": "claude-haiku-5-5"}
++{"work_id": 4211201992, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational multimedia development for English reading instruction; not CS research in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 4283813889, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.4, "rationale": "Game-theoretic sequential equilibria in signaling games; closest fit is game theory, which is not a listed category, so low confidence; could be algorithmic game theory or ml_theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4396572044, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Lightweight transformer architecture with contrastive learning for fault diagnosis; architecture contribution applied to bearing signals, though the application may suggest scientific or signal domain.", "model": "claude-haiku-5-5"}
++{"work_id": 4412353443, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.55, "rationale": "Remote-sensing mapping of urban villages from building footprints and imagery with random forest classification.", "model": "claude-haiku-5-5"}
++{"work_id": 7219753707, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Community service on teacher technology literacy; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2806158045, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Fuzzy logic controller comparison for a gantry crane; control engineering, closest to robotics but not clearly within the ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 2982518672, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Wireline communication system design using noise statistics; communications engineering, weakly fits computer_networks.", "model": "claude-haiku-5-5"}
++{"work_id": 7204116048, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Systematic review of digital educational platforms in primary education; education research, not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3112863481, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Multimedia network teaching platform for physical education; education application rather than CS contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3176660608, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Spectral unmixing of mineral mixtures; planetary science/remote sensing spectroscopy rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4206870115, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Non-CS educational/communication study on digital education in French; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4409833556, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Cultural information dissemination infographic design for cultural literacy; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2990710531, "categories": ["robotics"], "primary": "robotics", "confidence": 0.45, "rationale": "Control system for contactless small-ball air suspension in pipelines; control engineering closest to robotics/control, weakly CS-adjacent.", "model": "claude-haiku-5-5"}
++{"work_id": 4386952809, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "User survey on gaming motives and well-being; user study of player experience fits HCI.", "model": "claude-haiku-5-5"}
++{"work_id": 4225947117, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Qualitative study of school library management; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3039944072, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Engineering program accreditation and relevance study using viable systems model; educational management, not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4389567417, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Qualitative English pronunciation teaching study; education, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3109176755, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Combinatorics result on Kakeya sets over rings; pure mathematics with lower bounds on set size, closest to combinatorics/complexity-adjacent, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7128287921, "categories": [], "primary": null, "confidence": 0.7, "rationale": "AI in textile design education case study; educational research with AI tools, mainly not CS methodology contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3093363583, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Development of a mobile educational app for children; interactive learning tool, weakly HCI, mostly educational.", "model": "claude-haiku-5-5"}
++{"work_id": 3035934964, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.4, "rationale": "Mathematical optimization/complementarity theory on existence and uniqueness of solutions; closest ontology fit is theoretical, though it is more operations research/mathematics than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4399338544, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Radar cooperative tracking via weighted least squares; signal/sensor tracking not clearly a CS category, weakly fits robotics or none.", "model": "claude-haiku-5-5"}
++{"work_id": 4388262188, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Millimeter-wave radar displacement measurement is instrumentation and signal processing rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 2967228095, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Survey sampling estimator theory for population means; statistics, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4285385149, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Digital competence of computing teachers via DigCompEdu; education research, though software-education adjacent.", "model": "claude-haiku-5-5"}
++{"work_id": 4302287363, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Regional literacy campaign survey in education; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4393541793, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Archaeological visibility analysis of megaliths; archaeology using GIS, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4388202433, "categories": [], "primary": null, "confidence": 0.9, "rationale": "English listening pedagogy with multimedia; language education, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3200409248, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Neuroimaging study of encephalitis patients; clinical neuroscience, not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3007072699, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Automatic heart sound classification device for medical screening; biomedical informatics with audio signal processing, borderline between speech_audio and biomedical.", "model": "claude-haiku-5-5"}
++{"work_id": 3007925741, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Ultrasonic clamp-on flowmeter instrumentation; not a clear CS category, closest is sensing hardware, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4387771625, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education research on CS teacher stereotypes; not computer-science research in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 4297802412, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education policy analysis; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7139130213, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Modelling AS/RS cycle times with timed Petri nets for discrete event systems; closest is industrial automation, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4385550355, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Humanities analysis of a video game; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4409248967, "categories": ["computer_architecture"], "primary": null, "confidence": 0.25, "rationale": "Sensor array sampling layout for magnetocardiography; sensing instrumentation, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4394836987, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Pure mathematics on fractional differential equations; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3193522400, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Preschool education curriculum analysis; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3098057909, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Social science study of online communities in Africa; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4413096664, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Computer-supported collaborative learning applied to English education; educational research rather than CS category, weakly HCI.", "model": "claude-haiku-5-5"}
++{"work_id": 7166265487, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Course catalog description for a capstone course; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3197510543, "categories": ["database_systems", "information_retrieval"], "primary": null, "confidence": 0.2, "rationale": "Learning analytics intervention for online statistics dropout; not clearly within any CS ontology category, so no category assigned conservatively.", "model": "claude-haiku-5-5"}
++{"work_id": 4360976908, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational study of TikTok for English vocabulary learning; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7215297435, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational technology development for economics students; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4281735920, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Pedagogical article on presentation skills of graduates; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3006201666, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Android learning application development for vocational schools; educational application rather than a CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 4291313249, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Bibliographic discussion of distance education in Brazil; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7219459307, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Survey of e-learning resource use in business education; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2803291160, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Analysis of YouTube Zika videos and user responses; social media/health communication, closest to HCI-style social computing but not clearly CS-core, so left unassigned conservatively.", "model": "claude-haiku-5-5"}
++{"work_id": 4236001208, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "ARMA time-series models are statistical modelling; statistics chapter with weak fit to CS categories, assigned conservatively to probabilistic models.", "model": "claude-haiku-5-5"}
++{"work_id": 4402341510, "categories": ["ai_safety", "scientific_computing"], "primary": "ai_safety", "confidence": 0.1, "rationale": "Expert system for neurological diagnosis via Dempster-Shafer evidence theory; this is a classical AI/medical decision-support system, not clearly in any ontology category. Best fit is weak; likely computational_biology for health informatics, but the paper is closer to classical expert systems, so no confident category.", "model": "claude-haiku-5-5"}
++{"work_id": 2957642544, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational/applied linguistics study on mobile apps for EFL grammar learning; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4281729487, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Design of a children's batik drawing book; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4281615972, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Nursing education readiness survey on information technology use; educational research rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4415542380, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Systematic review on reading literacy and communication technology; communication/education studies, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3116015520, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Law education paper on web-based distance learning; pedagogy rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4396949355, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Characterization of random telegraph noise in FinFET transistors at cryogenic temperature; device-level semiconductor reliability, closest to computer architecture/hardware design.", "model": "claude-haiku-5-5"}
++{"work_id": 3201351346, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "CMOS RF energy-harvesting rectifier circuit design; low-power hardware/circuit design, closest to computer_architecture (VLSI and low-power design).", "model": "claude-haiku-5-5"}
++{"work_id": 2891299477, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Dielectric elastomer energy harvester analysis; materials and electromechanical engineering rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3087361831, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Chemical process control of an absorption column using PID and fuzzy logic; control engineering rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3081174150, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Russian paper on big data methods for e-learning; not clearly a CS category in the ontology, mostly education/general data processing; conservative assignment of none.", "model": "claude-haiku-5-5"}
++{"work_id": 3207299996, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Geoscience satellite observation of earthquake precursors; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4390807490, "categories": [], "primary": null, "confidence": 0.9, "rationale": "History of an IEEE journal's editorship; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3138970505, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education analytics of student platform logs; education research rather than a CS category, conservative none.", "model": "claude-haiku-5-5"}
++{"work_id": 3212108115, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Survey of students on internet project participation in education; not a clear CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4389301359, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Item response theory modeling of student persistence; statistical modeling of learners, borderline with ml_theory.", "model": "claude-haiku-5-5"}
++{"work_id": 3120906435, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Social science study of migration inference from geolocated tweets; data-methodology focus but not clearly CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4249995757, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Microgrid scheduling with particle swarm optimization; power systems application of metaheuristics, borderline algorithms but mainly energy engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 3159049567, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library management survey on internet services correlation; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3175870648, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library communication on Web 2.0 study; information science, not CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7133374282, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education research on digital competence of students; not a CS research contribution in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 2994564899, "categories": [], "primary": null, "confidence": 0.5, "rationale": "General education-system article on ICT use; not clearly computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4405519900, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Media studies of the Chilean telenovela industry; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4403910223, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Teacher professional development literature review on AI in education; education research rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4285808510, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Language learning mobile game in education; educational technology, not clearly CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4367555086, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Educational study on IoT applications in a personal learning environment for programming skills; education research primarily.", "model": "claude-haiku-5-5"}
++{"work_id": 4385356940, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Language education case study on online ELT during Covid-19; not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4377195986, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Control theory for Markov jump systems with event-triggered control; closest fit is a control/systems topic, weakly mapped to ml_theory; largely outside ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 4413172752, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational study of video media on student learning outcomes; not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4312554546, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Piezoelectric energy harvesting power converter circuit design; hardware/low-power circuits, closest to computer_architecture low-power design, though primarily electrical engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 2994665920, "categories": ["computer_architecture"], "primary": null, "confidence": 0.2, "rationale": "Soil moisture estimation via distributed temperature sensing is environmental sensing/geoscience instrumentation, not clearly a CS category; weak fit to sensors-based measurement at best, so conservatively assigned no category.", "model": "claude-haiku-5-5"}
++{"work_id": 3184466402, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education policy and digital educational resources comparison; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7203518940, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Higher education teaching session about AI adoption; pedagogy rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4298631416, "categories": [], "primary": null, "confidence": 0.85, "rationale": "E-learning effectiveness case study in higher education; educational management, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4402526116, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Information/data literacy scoping review; library and information science rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7169646709, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Essay on complexity science applied to organisations; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4280616584, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Documentary analysis of education 4.0 modality and regional development; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4394974551, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Legal scholarship on cybercrime and criminal law; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7201179105, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Model of uncertain reasoning chains with stochastic matrices resembles formal reasoning/probabilistic modelling, but the abstract is mainly decision-theoretic and conservatively not clearly in an ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 3152238796, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational psychology study on smartphone use and information retention in students; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7148607934, "categories": ["robotics", "computer_networks"], "primary": "robotics", "confidence": 0.35, "rationale": "Railway level-crossing safety with autonomous train transponder alerting vehicles; cyber-physical autonomy, but abstract is about safety context and is not clearly a CS contribution beyond transport systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4387754658, "categories": ["ml_theory"], "primary": null, "confidence": 0.1, "rationale": "Stability of variational inequalities in Hilbert spaces is pure mathematical optimisation, not computer science; conservative assignment of none.", "model": "claude-haiku-5-5"}
++{"work_id": 2943056810, "categories": ["deep_learning_architectures"], "primary": null, "confidence": 0.15, "rationale": "Coexistence of synchronization in coupled piecewise-constant oscillators is nonlinear circuit/dynamics analysis; closest is circuits, not a clear ontology match, so left minimal.", "model": "claude-haiku-5-5"}
++{"work_id": 4402215387, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Philosophical introduction to Edgar Morin's complexity thought; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3036170140, "categories": ["database_systems"], "primary": null, "confidence": 0.2, "rationale": "E-commerce information system design for a construction company; information systems development, not clearly any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 3005366057, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Generic practitioner tips on data automation projects with no clear CS research contribution in the abstract.", "model": "claude-haiku-5-5"}
++{"work_id": 3212504661, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Process sensor technology roadmap for process industry; engineering/instrumentation rather than a CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 7133133565, "categories": [], "primary": null, "confidence": 0.25, "rationale": "Essay on AI in education from a philosophy-of-education angle; not primarily CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3035309673, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Educational technology in higher education review; education research rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3205297713, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Commentary on standards-setting process for communications; policy and standards, not a CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 3159931479, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Organizational/management psychology study of employees' IM use and stress; not computer science research per ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 7166124027, "categories": ["sustainable_computing"], "primary": null, "confidence": 0.3, "rationale": "Solar PV potential estimation is energy engineering/geospatial analysis, not computer science; weak fit at most.", "model": "claude-haiku-5-5"}
++{"work_id": 7220147160, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Media/education studies on television series structure; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7133386248, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library repository network retrospective; library science, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 2889050697, "categories": ["ai_agents"], "primary": null, "confidence": 0.2, "rationale": "Mixed logical dynamical modelling and optimal scheduling of multi-energy systems; control/energy systems, weakly related to none of the CS categories.", "model": "claude-haiku-5-5"}
++{"work_id": 4390251033, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Appears to be a non-technical essay on symbols/language; no CS contribution evident.", "model": "claude-haiku-5-5"}
++{"work_id": 7215366142, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library science article on librarians' communication role; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4376526549, "categories": [], "primary": null, "confidence": 0.95, "rationale": "English language teaching material development; education, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4388721237, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Coordinated pricing for power-transportation networks with optimization algorithms; power systems/energy economics, not clearly a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4413905440, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Education policy and distance teacher training in Algeria; educational research, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4407389586, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Optical small-angle measurement via speckle pattern imaging with a camera; image-based measurement, not CS-core but closest to computer vision/metrology.", "model": "claude-haiku-5-5"}
++{"work_id": 4311440779, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Management/business study of digital transformation and environmental performance; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4322706772, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Electrical machine design for industrial electronics; electrical engineering, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4399577834, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational study on WebQuest for EFL writing; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4294309008, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Editorial-style piece on IT strategy consulting career with no abstract; not clearly CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4387450145, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Community training on Canva for entrepreneurship products; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2924637482, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Communication studies thesis on Instagram infographics; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4388692834, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Describes a consortium organization for information exchange among Pacific Rim universities; organizational description rather than CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4378804495, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Personality recognition from ECG during VR emotional stimuli; affective computing with VR user studies, with physiological signal processing component.", "model": "claude-haiku-5-5"}
++{"work_id": 2899278391, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "ML detection of QRS fragmentation from ECG for clinical biomarker; biomedical/health informatics method.", "model": "claude-haiku-5-5"}
++{"work_id": 7215205522, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Community service training on MS Excel report cards for teachers; not a CS research contribution in the ontology sense, so no category applied.", "model": "claude-haiku-5-5"}
++{"work_id": 2988437044, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Educational media development (R&D) for mathematics video learning; education research rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 3217539441, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Educational/learning perceptions study on non-cognitive skills; social-science style study, not clearly a CS category, though ICT-enabled.", "model": "claude-haiku-5-5"}
++{"work_id": 7172412385, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.3, "rationale": "Pure mathematics on Hurwitz quaternions and metacommutation permutations; closest fit is algebra/combinatorics, weakly matching no CS category, so low-confidence assignment kept minimal.", "model": "claude-haiku-5-5"}
++{"work_id": 4410790364, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Education management discussion on vocational experimental training bases; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4367663421, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Industrial engineering model for remanufactured product quality; uses neural network parameter estimation but the main contribution is engineering management, not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4413050241, "categories": ["deep_learning_architectures"], "primary": null, "confidence": 0.3, "rationale": "AI prediction models for preterm birth; primarily a clinical/biomedical informatics application, so applying computational_biology would be the closest fit but it is weak; conservatively this is a medical application of ML.", "model": "claude-haiku-5-5"}
++{"work_id": 3137938609, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Library and knowledge management study in Arabic education context; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7215521419, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Community service introducing Scratch programming to elementary school; educational outreach rather than a CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 2899602491, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Information systems study on technology supporting elderly patient adherence and motivation, a user-centred interactive systems question.", "model": "claude-haiku-5-5"}
++{"work_id": 4383695424, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Analytic number theory (universality of L-functions), mathematics rather than CS; not a CS research paper, so conservatively no category.", "model": "claude-haiku-5-5"}
++{"work_id": 2890545363, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Spatial ecological/epidemiological mathematical model; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 2921387549, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Education survey of expert perceptions on talent; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4411385469, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Airport security and profiling study in security studies; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4220744985, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Indonesian language education using animation; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4385809862, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Health informatics evaluation of data sharing in NHS; health services research rather than CS methodology, so no category assigned conservatively.", "model": "claude-haiku-5-5"}
++{"work_id": 4405194088, "categories": ["nlp"], "primary": "nlp", "confidence": 0.5, "rationale": "Uses NLP tools for lexical analysis of a historical Portuguese corpus; primarily digital humanities but computational linguistics framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4406402165, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Comparison of PV simulation software with solar plant data; energy engineering rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4235935708, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Mathematical model of nonlinear oscillations in a clock generator; applied mathematics/engineering, not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4410359890, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Economic and governance policy report on Africa; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4395448276, "categories": ["nlp"], "primary": null, "confidence": 0.1, "rationale": "Review of AI in higher education; not a CS research contribution in the ontology's sense, so essentially no category applies; very low-confidence fallback.", "model": "claude-haiku-5-5"}
++{"work_id": 7127568039, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational/health-promotion survey of university students' digital teaching; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3034138045, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational administration study on knowledge engineering in schools; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4381193254, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Philosophy of science/psychology on the term repressed memory; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7220079635, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Language-learning application study using AI in apps; education focus rather than CS contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3213996293, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Optical SPR sensor physics/sensing instrumentation; not computer science category.", "model": "claude-haiku-5-5"}
++{"work_id": 2990253946, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.4, "rationale": "IoT-based evacuation support with evolutionary deployment algorithm; closest fit is IoT/edge-style system, though also related to computer_networks.", "model": "claude-haiku-5-5"}
++{"work_id": 7203847693, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Edited volume on technology in language education; educational research rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4413000914, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Survey-based psychology/education study on AI literacy and teacher burnout; not CS contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 2997518615, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational metasynthesis on digital competencies; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7221210578, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Community service/dissemination report on online learning in Indonesian universities; not a CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4405738901, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Mathematical modelling of neuron networks with fractional Laplacian and reaction-diffusion equations; closest to computational modelling, with a neuroscience flavour.", "model": "claude-haiku-5-5"}
++{"work_id": 4210484898, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Intelligent tool holder with real-time cutting force sensing for drilling; manufacturing/sensing engineering, weakly fits embedded sensing hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4382986001, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Review of article processing charges in scholarly publishing; bibliometrics/science policy, not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 2946426731, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Study of prosumers in crowdfunded film industry; media and business studies rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4386047869, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Machine learning on hyperspectral imaging features for non-invasive food quality inspection; imaging-based analysis with industrial application.", "model": "claude-haiku-5-5"}
++{"work_id": 7216070232, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Editorial on ethics and nutrition profession; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4387426085, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education study on talent cultivation for pesticide residue detection; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3168268628, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.3, "rationale": "Open-source web tool for group formation using forgetting curves for collaboration; closest fit is recommendation-style grouping, but also CSCW-related (hci).", "model": "claude-haiku-5-5"}
++{"work_id": 4405513290, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Synchronization in adaptive simplicial complexes is nonlinear dynamics/network science physics, not a clear CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4390074168, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Management/information systems study of digital transformation in manufacturing; not a CS ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 2947916735, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Knowledge diagnostics testing model with doubt measurement for e-learning systems; closest to user-facing interactive systems, weakly assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 7213988930, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education and library science study of school libraries; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3112250358, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Solar resource estimation from meteorological and satellite data; energy/geoscience application rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7148627082, "categories": ["deep_learning_architectures", "computer_vision"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Bio-inspired mixed-variable optimisation (mixed GWO) applied to SVM image classification and multispectral image denoising/unmixing.", "model": "claude-haiku-5-5"}
++{"work_id": 7218427777, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library service management study in a madrasah; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4352989561, "categories": ["database_systems"], "primary": null, "confidence": 0.3, "rationale": "Information dissemination model in online social networks; social network analysis, no clear fit among categories, so kept conservative.", "model": "claude-haiku-5-5"}
++{"work_id": 7128218809, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Language education study on messaging apps for French learning; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4406799739, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Pure mathematics on variational inequalities in Banach spaces; not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 3129513804, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Essay on computing education research; education topic rather than a CS technical category.", "model": "claude-haiku-5-5"}
++{"work_id": 2915662082, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Computing education/academic workforce partnership; not a CS research contribution fitting any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 3006452720, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.45, "rationale": "Spatiotemporal missing-data imputation for geospatial data; closest to data management/cleaning, though it is also data-analysis methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4388634849, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Social science study of women's abuse on social media in ICT4D; HCI-adjacent but primarily qualitative social research, so conservatively none.", "model": "claude-haiku-5-5"}
++{"work_id": 4205387857, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Time-series regression of climate variables; statistical/environmental science, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7162195214, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Random forest regression applied to educational outcomes; educational data analysis rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7112126402, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Overview of AI use in education in Russian; applied educational discussion without a clear CS research contribution category.", "model": "claude-haiku-5-5"}
++{"work_id": 3212702011, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Photodiode characterization for optical communication; photonics/device physics, not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4226049650, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Optics textbook chapter on interference and diffraction; physics, not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4409157433, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Financial network analysis via local Gaussian correlation; econophysics/statistics rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 3080490474, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Underwater shearography nondestructive testing; optical engineering, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4411680001, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "AI tools in children's creative art activities; education-focused, weakly fits HCI for interactive creative tools, not clearly CS core.", "model": "claude-haiku-5-5"}
++{"work_id": 4312292089, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Qualitative study of an after-school programme for child development; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4412583047, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Educational program evaluation in Indonesian primary schools; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4391250710, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Online engineering laboratory teaching approaches; engineering education, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2954250517, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library work order management workshop in Alma; library science, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3129713058, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Educational technology experiment on e-learning strategy for teaching C++; educational research rather than CS contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3206918036, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.4, "rationale": "Quality assessment and validation of satellite land products from Copernicus; remote sensing, closest to scientific or vision data analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4412965306, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "LSTM-BiLSTM model applied to infrastructure health monitoring; application of deep sequence models, could also fit scientific_computing.", "model": "claude-haiku-5-5"}
++{"work_id": 4415671687, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Geostatistics-enhanced supervised classification with uncertainty quantification; ML methodology applied to geoscience.", "model": "claude-haiku-5-5"}
++{"work_id": 4404412038, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Career trajectory analysis of executives using LinkedIn data; information systems management study, not core CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3202598135, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Spectrum allocation in wireless networks optimized with a bio-inspired metaheuristic; networking application of an optimization algorithm.", "model": "claude-haiku-5-5"}
++{"work_id": 2969566816, "categories": ["computer_architecture", "cybersecurity", "computer_networks"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "FPGA-based OpenFlow switch hardware architecture with DDoS protection functions.", "model": "claude-haiku-5-5"}
++{"work_id": 4411949771, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.9, "rationale": "RPL routing protocol extension for IoT low-power lossy networks with QoS.", "model": "claude-haiku-5-5"}
++{"work_id": 4409370970, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Transformer-based intrusion detection for DDoS attacks on 5G networks.", "model": "claude-haiku-5-5"}
++{"work_id": 7117461702, "categories": ["computer_networks", "privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Privacy-preserving data management and compression in wireless sensor networks.", "model": "claude-haiku-5-5"}
++{"work_id": 2890181437, "categories": ["computer_networks", "cybersecurity"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Satellite telecommand link availability with defences against jamming in CCSDS protocols.", "model": "claude-haiku-5-5"}
++{"work_id": 4399801188, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.85, "rationale": "Performance analysis of fluid antenna backscatter communication systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7162436281, "categories": ["cybersecurity", "hci"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Replay-resistant offline authentication for BLE smart lock using signed credentials.", "model": "claude-haiku-5-5"}
++{"work_id": 7204457438, "categories": ["cybersecurity", "computer_networks", "ml_theory"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Zero Trust security framework with Random Forest detection of malicious nodes in 6G IoT.", "model": "claude-haiku-5-5"}
++{"work_id": 2982187970, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.9, "rationale": "UAV-aided multi-hop backhaul scheduling and path planning in RAN.", "model": "claude-haiku-5-5"}
++{"work_id": 3011142689, "categories": ["computer_networks", "cloud_computing"], "primary": "computer_networks", "confidence": 0.6, "rationale": "5G digital enterprise, SDN/NFV and synchronisation network service for neutral hosts; networking-focused platform paper.", "model": "claude-haiku-5-5"}
++{"work_id": 4413557542, "categories": ["computer_networks", "robotics"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Analytical latency model of DDS reliable delivery over UDP; networking performance, applied to ROS 2 middleware.", "model": "claude-haiku-5-5"}
++{"work_id": 2887023911, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Broadcast transmission scheme (LDM/BST-OFDM, LDPC) for digital terrestrial broadcasting; communications/broadcast system design, closest to networks.", "model": "claude-haiku-5-5"}
++{"work_id": 7217270607, "categories": ["cybersecurity", "ai_agents"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "AI-augmented network security and threat detection evaluation in enterprise networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4414744692, "categories": ["computer_networks", "federated_learning"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Learned digital over-the-air aggregation for federated edge learning; wireless communication with FL as application.", "model": "claude-haiku-5-5"}
++{"work_id": 3202853167, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "ARQ-based broadcast retransmission and index coding analysis for wireless networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4400315068, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "ML-based anomaly detection for IoT network security.", "model": "claude-haiku-5-5"}
++{"work_id": 7138862310, "categories": ["reinforcement_learning", "computer_networks"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Novel distributional offline RL algorithm applied to bandwidth estimation for video conferencing QoE.", "model": "claude-haiku-5-5"}
++{"work_id": 4382248155, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "In-band network telemetry and data plane programmability for network monitoring and failure recovery.", "model": "claude-haiku-5-5"}
++{"work_id": 3137488241, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Security analysis of WPA3 transition mode and downgrade attacks on Wi-Fi.", "model": "claude-haiku-5-5"}
++{"work_id": 4291165406, "categories": ["computer_networks", "computer_vision"], "primary": null, "confidence": 0.3, "rationale": "Mobile sensor network for air pollution monitoring; environmental sensing, not clearly CS category. Sensor network framing suggests computer_networks weakly; not clearly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4400276720, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "System-level simulation of mobility and beamforming in 5G FR2 networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4417252163, "categories": ["computer_networks", "deep_learning_architectures"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Channel estimation for massive MIMO using diffusion generative models; wireless communications with generative model contribution, possibly generative_ai.", "model": "claude-haiku-5-5"}
++{"work_id": 3157147353, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.85, "rationale": "Scheduling scheme for 802.11 WLAN congestion and TXOP allocation.", "model": "claude-haiku-5-5"}
++{"work_id": 4402187505, "categories": ["computer_networks", "nlp"], "primary": "computer_networks", "confidence": 0.5, "rationale": "LSTM with CRF for intent-driven routing in power backbone networks; routing is networking.", "model": "claude-haiku-5-5"}
++{"work_id": 7112597741, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Study of DNSSEC and a DNS cache poisoning attack in simulation.", "model": "claude-haiku-5-5"}
++{"work_id": 3039350537, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Energy-aware transmission scheduling for UAV relay networks.", "model": "claude-haiku-5-5"}
++{"work_id": 7202339439, "categories": ["cybersecurity", "ml_theory"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Detecting spectrum sensing data falsification attacks with classifiers in cognitive radio networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4385988957, "categories": ["computer_networks", "reinforcement_learning"], "primary": "computer_networks", "confidence": 0.7, "rationale": "TSN routing and scheduling using distributional RL; networking focus with RL method.", "model": "claude-haiku-5-5"}
++{"work_id": 3160907559, "categories": ["edge_computing"], "primary": null, "confidence": 0.4, "rationale": "Systematic review of IoT architecture challenges; IoT architecture is loosely related to edge computing but topic is broad.", "model": "claude-haiku-5-5"}
++{"work_id": 4290991003, "categories": ["generative_ai", "computer_networks"], "primary": "generative_ai", "confidence": 0.8, "rationale": "GAN-based synthesis of network header traces, applied to networking tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 4210382415, "categories": ["distributed_computing", "ml_theory"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Coded distributed multi-task learning to reduce communication load across users and server.", "model": "claude-haiku-5-5"}
++{"work_id": 4375957559, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "STAR-RIS-assisted covert communication with beamforming optimization for wireless networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4404627904, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Data-driven blind signal separation and target signal extraction using multimodal clues for co-channel interference; signal separation of audio-like signals.", "model": "claude-haiku-5-5"}
++{"work_id": 4389101774, "categories": ["hci", "computer_networks"], "primary": "hci", "confidence": 0.5, "rationale": "Wearable wireless breath monitoring system with epidermal sensor, applied to users in motion.", "model": "claude-haiku-5-5"}
++{"work_id": 4416894542, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Conceptual network architecture framework using SDN/NFV for internet performance.", "model": "claude-haiku-5-5"}
++{"work_id": 2775587204, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational survey study on social networking sites in education; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3013810395, "categories": ["edge_computing", "reinforcement_learning"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Vehicular edge computing data scheduling with computation offloading, solved via deep Q-network.", "model": "claude-haiku-5-5"}
++{"work_id": 4386918894, "categories": ["reinforcement_learning", "computer_networks"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Multi-agent deep deterministic policy gradient for RAN slicing in vehicular networks.", "model": "claude-haiku-5-5"}
++{"work_id": 2806094603, "categories": ["computer_networks", "ai_accelerators"], "primary": "computer_networks", "confidence": 0.6, "rationale": "GPU-accelerated SVM for network traffic classification, a networking application with hardware acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 3207037104, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Survey of attacks and countermeasures in wireless sensor networks; security is the main focus, with networking context.", "model": "claude-haiku-5-5"}
++{"work_id": 4408441962, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Attention-based neural network for malicious IoT traffic detection; intrusion detection is the application, with an architectural element.", "model": "claude-haiku-5-5"}
++{"work_id": 2922268566, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Editorial on SDN, IoT security, SDWAN and cloud networking; broad networking overview.", "model": "claude-haiku-5-5"}
++{"work_id": 2921781262, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Reduces duplicate OpenFlow requests to SDN controller; SDN network architecture contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7107949444, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Performance measurement of a campus Wi-Fi network; wireless networking evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 3176443371, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.55, "rationale": "Cache updating for age of information with Markov decision optimisation; caching in networks with optimisation methods.", "model": "claude-haiku-5-5"}
++{"work_id": 3115846872, "categories": ["computer_networks", "robotics"], "primary": "computer_networks", "confidence": 0.5, "rationale": "V2X data dissemination for 3D-map aided GNSS positioning in vehicles; communication-centric with intelligent transport localisation.", "model": "claude-haiku-5-5"}
++{"work_id": 4381805382, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Survey of Sybil attack countermeasures in underwater sensor networks; security focus.", "model": "claude-haiku-5-5"}
++{"work_id": 2800766972, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.75, "rationale": "Reduces OpenFlow rule-change latency in data centre networks; SDN networking performance.", "model": "claude-haiku-5-5"}
++{"work_id": 4410810312, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Outage probability analysis of RIS-assisted multihop wireless networks; networking performance analysis with a deep learning estimator.", "model": "claude-haiku-5-5"}
++{"work_id": 7220552095, "categories": ["nlp", "hci"], "primary": "nlp", "confidence": 0.6, "rationale": "Text mining and sentiment analysis of reviews with TF-IDF and LDA; NLP-centred, with a tourism user-experience angle.", "model": "claude-haiku-5-5"}
++{"work_id": 4283810249, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Text span extraction and question generation for art passages; classical NLP for a low-resource domain.", "model": "claude-haiku-5-5"}
++{"work_id": 2909157681, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Semantic relationships between folksonomy tags via embeddings and community detection; closer to tagging/recommendation-style retrieval, with NLP-style embeddings.", "model": "claude-haiku-5-5"}
++{"work_id": 7124304832, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "POS tagging comparison of classical ML models for a low-resource language.", "model": "claude-haiku-5-5"}
++{"work_id": 4221152634, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Dataset for joint German text summarization and simplification; NLP resource.", "model": "claude-haiku-5-5"}
++{"work_id": 7160969555, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Czech valency lexicon resource for NLP applications; computational linguistics resource.", "model": "claude-haiku-5-5"}
++{"work_id": 3033351879, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.55, "rationale": "Modified transformer for task-oriented dialogue; dialogue is nlp, architecture modifications are deep learning.", "model": "claude-haiku-5-5"}
++{"work_id": 4353004366, "categories": ["nlp", "llm", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.6, "rationale": "Long-text understanding by reusing short-context pretrained language models; NLP evaluation on SCROLLS with transformer LMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4399355922, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.55, "rationale": "Document matching for chatbots comparing cosine similarity and BERT; retrieval-focused.", "model": "claude-haiku-5-5"}
++{"work_id": 4399657664, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Benchmark for entity summarization over knowledge graphs; combines graph algorithms and NLP, with evaluation resource framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7167593003, "categories": ["llm", "ai_safety", "nlp"], "primary": "llm", "confidence": 0.6, "rationale": "Mixture-of-experts framework on a shared LLM backbone for social media depression detection; LLM-centred clinical NLP classification, with ML methodology (LUPI) and mental health application.", "model": "claude-haiku-5-5"}
++{"work_id": 4387323905, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.8, "rationale": "Continued pretraining of T5 LLM families for table tasks with scaling to 11B; LLM pretraining and evaluation with table QA/semantic parsing framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7213562850, "categories": ["nlp", "scientific_computing"], "primary": "nlp", "confidence": 0.45, "rationale": "Lexical/semantic text representation model for landscape classification from vegetation descriptions; text processing with geographic application.", "model": "claude-haiku-5-5"}
++{"work_id": 7166888984, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.5, "rationale": "Attribution framework for diagram question answering, a vision-language task; abstract is mostly author metadata so confidence is limited.", "model": "claude-haiku-5-5"}
++{"work_id": 7203972941, "categories": ["nlp"], "primary": "nlp", "confidence": 0.4, "rationale": "BERT-based Arabic fake news sentiment analysis, but the abstract is conference boilerplate and the topic is inferred from the title.", "model": "claude-haiku-5-5"}
++{"work_id": 7166865552, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.7, "rationale": "Fine-tuning GPT-4.1-mini and prompting strategies for abductive event reasoning in a SemEval task; LLM-centred evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 7212985666, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.7, "rationale": "Evaluates LLMs for courtroom and check interpreting, including LLM-as-a-Judge; LLM machine translation evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 4414404922, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Large-scale multilingual speech dataset and pipeline for speech generation; speech synthesis resource.", "model": "claude-haiku-5-5"}
++{"work_id": 4377372007, "categories": ["llm", "ai_agents"], "primary": "llm", "confidence": 0.75, "rationale": "Framework for LLM self-correction using external tools; LLM reasoning and tool interaction, also agentic in style.", "model": "claude-haiku-5-5"}
++{"work_id": 4390215039, "categories": ["nlp", "formal_verification"], "primary": "nlp", "confidence": 0.5, "rationale": "Compositional semantics for interrogatives in CCG with a QA benchmark; computational linguistics and semantic parsing.", "model": "claude-haiku-5-5"}
++{"work_id": 2964776615, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Question generation from passages is a classical NLP generation task using encoder-decoder models, not LLM-centred.", "model": "claude-haiku-5-5"}
++{"work_id": 4412888626, "categories": ["llm", "ai_agents"], "primary": "llm", "confidence": 0.7, "rationale": "Studies LLaMA models' internal representations of Theory of Mind and alignment in conversational agents.", "model": "claude-haiku-5-5"}
++{"work_id": 4389523964, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Feature-based knowledge distillation for compressing neural network language models; a compression method as such.", "model": "claude-haiku-5-5"}
++{"work_id": 3035035809, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Box embeddings for representing hierarchies are a representation learning method; closest fit is ml_theory, though knowledge-graph focused.", "model": "claude-haiku-5-5"}
++{"work_id": 4285254984, "categories": ["hci"], "primary": null, "confidence": 0.3, "rationale": "Ontology-based digital transformation of railway education; not clearly within any core CS category, weakly fits software or hci education tooling.", "model": "claude-haiku-5-5"}
++{"work_id": 4396653207, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Thesis on sentence embeddings and linguistic structure in neural models, a computational linguistics topic.", "model": "claude-haiku-5-5"}
++{"work_id": 347871603, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Evaluation framework and resources for machine translation, an NLP evaluation contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4417356542, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.7, "rationale": "Benchmarks CNNs and prompt-tuned BERT for sentiment analysis, a classical NLP text classification task.", "model": "claude-haiku-5-5"}
++{"work_id": 4361006333, "categories": ["ml_theory", "computer_architecture"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Genetic algorithm with supervised classifiers for failure prediction in semiconductor manufacturing; applied ML rather than a clear category, with industrial focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4212987749, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "Fine-grained sentiment analysis of movie reviews with classical supervised classifiers, a text classification task.", "model": "claude-haiku-5-5"}
++{"work_id": 4387389341, "categories": ["nlp", "computational_biology"], "primary": "nlp", "confidence": 0.5, "rationale": "Machine learning on Japanese interview text to identify latent needs; text classification in a health-informatics context.", "model": "claude-haiku-5-5"}
++{"work_id": 2997143856, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.7, "rationale": "RNN-based sentiment classification of reviews, a classical NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 4396695606, "categories": ["deep_learning_architectures", "nlp"], "primary": "deep_learning_architectures", "confidence": 0.55, "rationale": "Noise-induced regularization for LSTM/BiLSTM models, evaluated on NLP tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 4406975371, "categories": ["llm"], "primary": "llm", "confidence": 0.6, "rationale": "Context-aware mechanism for large language model architectures and text generation.", "model": "claude-haiku-5-5"}
++{"work_id": 4404782050, "categories": ["llm", "reinforcement_learning"], "primary": "llm", "confidence": 0.75, "rationale": "Inference-time alignment of large language models using value guidance during decoding.", "model": "claude-haiku-5-5"}
++{"work_id": 7213318585, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Book on sentiment analysis spanning classical and deep learning NLP methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4200154996, "categories": ["nlp", "information_retrieval"], "primary": "nlp", "confidence": 0.5, "rationale": "Sentiment and topic classification of COVID-19 news text; primarily text mining.", "model": "claude-haiku-5-5"}
++{"work_id": 4225411551, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Structured sentiment analysis shared-task system using BART generation for opinion extraction.", "model": "claude-haiku-5-5"}
++{"work_id": 4403127614, "categories": ["nlp", "hci"], "primary": "nlp", "confidence": 0.5, "rationale": "Chatbot for e-commerce customer service built with NLP and ML; application-oriented.", "model": "claude-haiku-5-5"}
++{"work_id": 2971319154, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Erroneous data generation for neural grammatical error correction, an NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 4226489197, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Low-resource named entity recognition method, a classical NLP information extraction task.", "model": "claude-haiku-5-5"}
++{"work_id": 4285298945, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Abusive comment classification in Tamil is a text classification NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 4399923297, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Implementation of a personality-based chatbot integrated with web, studied via user experience; borderline between hci and nlp.", "model": "claude-haiku-5-5"}
++{"work_id": 4213262540, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Sentiment analysis and topic modeling of English and Tamil tweets.", "model": "claude-haiku-5-5"}
++{"work_id": 4405737917, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Naive Bayes sentiment classification of tweets about a presidential candidate.", "model": "claude-haiku-5-5"}
++{"work_id": 7169949244, "categories": ["nlp"], "primary": "nlp", "confidence": 0.5, "rationale": "Sentiment analysis of YouTube comments on fashion sustainability; abstract is only a title-like description.", "model": "claude-haiku-5-5"}
++{"work_id": 7217753392, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Review of embedding fusion methods for NLP and machine learning.", "model": "claude-haiku-5-5"}
++{"work_id": 4400512809, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "Sentiment and emotion analysis of X data on school reopening using supervised ML.", "model": "claude-haiku-5-5"}
++{"work_id": 7204475324, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Qualitative interview study of students' AI use in higher education; social science rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4408546512, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Recurrent neural network based single-document summarisation, a classical NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 2964302064, "categories": ["hci", "computer_vision"], "primary": "hci", "confidence": 0.5, "rationale": "EEG/MEG fusion for motor-imagery BCI is a neurotechnology signal-classification study; BCIs are an interactive-systems topic but the work is signal processing, so it is borderline between hci and computational_biology/ml methods. Chosen hci as the BCI interaction framing.", "model": "claude-haiku-5-5"}
++{"work_id": 2946097570, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "User-centred requirements elicitation with stakeholders for learning analytics dashboards is a participatory design study.", "model": "claude-haiku-5-5"}
++{"work_id": 3127843396, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Dissertation on interaction models for mixed-reality interactive experiences, user engagement and narratively charged objects.", "model": "claude-haiku-5-5"}
++{"work_id": 7140640315, "categories": ["computer_vision", "multimodal_ai", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Skeleton and multimodal human action recognition with graph convolution on NTU RGB+D; primarily video/pose recognition with multimodal fusion.", "model": "claude-haiku-5-5"}
++{"work_id": 7131851078, "categories": ["hci", "reinforcement_learning"], "primary": "hci", "confidence": 0.4, "rationale": "Educational framework with XAI, holographic displays and gamified learning for physical education; interactive learning system, with RL-driven gamification mentioned only in passing.", "model": "claude-haiku-5-5"}
++{"work_id": 3196036446, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Assets-based community engagement and design of socio-technical systems with a vulnerable population; HCI/CSCW fieldwork.", "model": "claude-haiku-5-5"}
++{"work_id": 2948072807, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Microwave reflection-type phase shifter circuit design; hardware/circuit topic that fits closest to computer_architecture's VLSI/low-power/EDA scope, though it is really RF electronics outside the ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 3039601072, "categories": ["hci"], "primary": "hci", "confidence": 0.85, "rationale": "Field study on how audio supports reminiscence for people living with blindness; accessibility and user experience.", "model": "claude-haiku-5-5"}
++{"work_id": 7171785185, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Systematic review of eye-tracking in infants with motor disabilities; clinical/health research rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4312560958, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Development and evaluation of augmented-reality edutainment media for mathematics learning; educational technology with user validation, closest to hci.", "model": "claude-haiku-5-5"}
++{"work_id": 4403160367, "categories": ["reinforcement_learning", "hci", "computer_graphics"], "primary": "hci", "confidence": 0.5, "rationale": "Exploratory agent evaluating procedurally generated game levels; spans game design evaluation (computer graphics/game environments) and agent-based exploration; HCI-leaning as player engagement evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 4312735085, "categories": ["hci"], "primary": "hci", "confidence": 0.9, "rationale": "User experience evaluation method with children, user study.", "model": "claude-haiku-5-5"}
++{"work_id": 7135288805, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Sociolinguistics study of Facebook's effect on writing; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4412960450, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Survey of student perception of augmented reality technology; user perception study of AR, education context.", "model": "claude-haiku-5-5"}
++{"work_id": 7220333334, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Design criteria for B2C e-commerce websites; web design and usability, borderline with information systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4409620177, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Survey of students' perceptions of an English-learning platform; educational research rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7203798403, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Conceptual framework on UX design in VR training systems, usability and presence.", "model": "claude-haiku-5-5"}
++{"work_id": 2980067317, "categories": ["speech_audio", "ai_safety"], "primary": "speech_audio", "confidence": 0.1, "rationale": "EEG motor imagery BCI signal classification; signal processing with SVM; closest fit is not clear, low confidence. Better fit might be computational biology or none.", "model": "claude-haiku-5-5"}
++{"work_id": 3166564139, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Web-based information system for practical work data management; information system design.", "model": "claude-haiku-5-5"}
++{"work_id": 4413023786, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Systematic review of mobile learning application design principles and learning outcomes; educational technology framed as HCI design.", "model": "claude-haiku-5-5"}
++{"work_id": 4317438351, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "VR educational game for process safety; interactive system for learners, primarily an HCI/VR application.", "model": "claude-haiku-5-5"}
++{"work_id": 4403626708, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Review of AR consumer behaviour and marketing; human-computer interaction from consumer perspective, not CS-core but closest fit.", "model": "claude-haiku-5-5"}
++{"work_id": 2951084085, "categories": ["distributed_computing", "robotics"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Dataflow processing with Apache Beam for scalable clinical sensor analytics; scaling-out system focus, with health-sensing application.", "model": "claude-haiku-5-5"}
++{"work_id": 7131915600, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Development of a frontend framework-based interface with microservices following agile methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4417220792, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Critical analysis of behaviour change design between HCI and behavioural science.", "model": "claude-haiku-5-5"}
++{"work_id": 7198871956, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Qualitative sociolinguistics study of text-speak; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7202300623, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Hybrid 2D-3D VR visualization with user study; interactive visualization systems for users.", "model": "claude-haiku-5-5"}
++{"work_id": 2790900283, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Framework for immersive virtual cognitive experiments with EEG and user cognition; HCI-oriented with some geographic information focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4221152344, "categories": ["hci", "ai_safety"], "primary": "hci", "confidence": 0.6, "rationale": "User study of human-AI teamwork and bias in hiring decisions; human-AI interaction with fairness concerns.", "model": "claude-haiku-5-5"}
++{"work_id": 7119905577, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Empirical study of presence and immersive experience in mixed reality cultural tourism.", "model": "claude-haiku-5-5"}
++{"work_id": 4405928278, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Systematic review of metaverse research in business contexts; closest to human-computer interaction and user experience themes, though it is a literature review on business applications.", "model": "claude-haiku-5-5"}
++{"work_id": 3186159243, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Pilot study of VR use in classroom teaching examining student motivation and teacher perspectives; interactive systems and users.", "model": "claude-haiku-5-5"}
++{"work_id": 7112876774, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Study of XR headsets for code tracing with eye gaze and motivation measures; user study of interactive systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4225091482, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Open-ended sensor kit for office workers' self-tracking; personal informatics and user studies.", "model": "claude-haiku-5-5"}
++{"work_id": 2911964379, "categories": ["reinforcement_learning", "information_retrieval"], "primary": "reinforcement_learning", "confidence": 0.4, "rationale": "Uses sequential zero-determinant game-theoretic strategies to incentivise crowd workers; game-theory based incentives tied to crowdsourcing quality control.", "model": "claude-haiku-5-5"}
++{"work_id": 4285312412, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Cultural and media studies essay on TikTok attention and shock; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4413901110, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education survey on student perceptions of AI in language classes; primarily education research rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7123361757, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Affective science study of emotional responses to VR 360-degree video; psychology rather than computer science, though it touches VR.", "model": "claude-haiku-5-5"}
++{"work_id": 3216674623, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Wearable assistive communication device for deaf-blind users; accessibility and assistive technology.", "model": "claude-haiku-5-5"}
++{"work_id": 7161768941, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Sociolinguistic study of address terms on Twitter/X; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4413188996, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Essay on digital divide in teacher education; education research rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3189825209, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Human-centered design research and critical design communication of cultural insights with user studies.", "model": "claude-haiku-5-5"}
++{"work_id": 4400638471, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "User-centered design guide for low-tech artefacts, cognitive ergonomics focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3138774926, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Empirical study of eye and mouse dynamics during web browsing with 151 participants.", "model": "claude-haiku-5-5"}
++{"work_id": 7216627434, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Community service training on AR math learning media for teachers; education outreach rather than CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4400230898, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Wireless intracortical BCI chip design with on-chip spike detection and compression; hardware system design, possibly biomedical.", "model": "claude-haiku-5-5"}
++{"work_id": 7124922965, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Position paper on annotation as interaction technique for tactile graphics and accessibility for blind users.", "model": "claude-haiku-5-5"}
++{"work_id": 4200159390, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Overview of systematic reviews on digital interventions for mental health; technology-focused health informatics with HCI framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7125427419, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Thematic review and guidelines for incentive design in human-AI decision-making user studies.", "model": "claude-haiku-5-5"}
++{"work_id": 7132825323, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Survey experiment on consumer responses to hyper-realistic digital humans, human-AI interaction and perception; marketing-oriented.", "model": "claude-haiku-5-5"}
++{"work_id": 7117881138, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Adaptive weight aggregation in federated learning applied to polyp segmentation, a medical image analysis task.", "model": "claude-haiku-5-5"}
++{"work_id": 4317566642, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Moment-based joint density estimation for synthetic micro-data; statistical methodology with disclosure protection motivation, closer to statistics than privacy computation per se.", "model": "claude-haiku-5-5"}
++{"work_id": 4388516975, "categories": ["privacy_preserving_computation", "computer_vision"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "Homomorphic encryption applied to CT image classification for privacy-preserving lung cancer diagnosis.", "model": "claude-haiku-5-5"}
++{"work_id": 7110219066, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.85, "rationale": "Differential privacy noise distributions for smooth sensitivity mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 7123282485, "categories": ["ai_safety", "privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Re-identification attack on released interview data using agentic LLMs; privacy attack with LLM agents.", "model": "claude-haiku-5-5"}
++{"work_id": 4313598458, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Local differential privacy mechanisms with causal inference and minimax analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4360995097, "categories": ["federated_learning", "privacy_preserving_computation", "cybersecurity"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Gradient inversion privacy attack in federated learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7171902819, "categories": ["privacy_preserving_computation", "hci"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Differential privacy for EEG data via RF perturbation; sensing and privacy.", "model": "claude-haiku-5-5"}
++{"work_id": 4405184264, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Privacy-utility analysis of the Topics API with differential privacy bounds.", "model": "claude-haiku-5-5"}
++{"work_id": 4404037745, "categories": ["computational_complexity", "applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Information-theoretic symmetric private information retrieval capacity over noisy channels.", "model": "claude-haiku-5-5"}
++{"work_id": 3139240920, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.9, "rationale": "Proof-of-concept k-anonymity data anonymization, a privacy technique for data release.", "model": "claude-haiku-5-5"}
++{"work_id": 4226227596, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Privacy rating of mobile apps via machine learning regression on permissions; privacy-focused, with an ML prediction component.", "model": "claude-haiku-5-5"}
++{"work_id": 4318812474, "categories": ["privacy_preserving_computation", "federated_learning", "deep_learning_architectures"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "Privacy-preserving split learning with patch shuffling on transformers, defending against attacks on training data.", "model": "claude-haiku-5-5"}
++{"work_id": 4411949626, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Federated learning under domain heterogeneity with prototype contrastive learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7214618239, "categories": ["cybersecurity", "nlp", "generative_ai", "ai_safety"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Adversarial fraud transaction generation to evaluate fraud detection systems; fraud security with LLM-based data generation and catastrophic forgetting.", "model": "claude-haiku-5-5"}
++{"work_id": 4405710613, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.5, "rationale": "Book on graph learning with emphasis on privacy preservation; graph learning is also central.", "model": "claude-haiku-5-5"}
++{"work_id": 7165521315, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Mechanism design and optimal pricing for data broker; economic theory rather than clearly a CS category, closest to ml_theory-style principled analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 3137146239, "categories": ["privacy_preserving_computation", "computer_networks"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Privacy-preserving BLE identity broadcast for contact tracing using secret sharing; wireless networking and privacy.", "model": "claude-haiku-5-5"}
++{"work_id": 2971047472, "categories": ["privacy_preserving_computation", "computational_complexity"], "primary": "privacy_preserving_computation", "confidence": 0.65, "rationale": "Approximate differential privacy in labelled Markov chains; complexity results and bisimilarity distances.", "model": "claude-haiku-5-5"}
++{"work_id": 7168322783, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Federated few-shot learning for AFib ECG detection across institutions.", "model": "claude-haiku-5-5"}
++{"work_id": 4393952232, "categories": ["privacy_preserving_computation", "robotics"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Citizen privacy for delivery drone video using secure two-party computation and audit trails; drones are robotic systems but the focus is privacy.", "model": "claude-haiku-5-5"}
++{"work_id": 7133196926, "categories": ["privacy_preserving_computation", "deep_learning_architectures"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Differentially private training (DP-SGD variants) for large embedding models preserving gradient sparsity.", "model": "claude-haiku-5-5"}
++{"work_id": 2899391355, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Empirical study of GDPR data portability requests and compliance; privacy regulation focus.", "model": "claude-haiku-5-5"}
++{"work_id": 2945715100, "categories": ["privacy_preserving_computation", "hci"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Privacy system for smart service platforms limiting disclosed information in IoT environments.", "model": "claude-haiku-5-5"}
++{"work_id": 2886835905, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Location privacy anonymisation scheme for continuous location-based queries.", "model": "claude-haiku-5-5"}
++{"work_id": 7125647931, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Ukrainian-language paper on AI in public finance and economic governance; not clearly computer-science research, mainly economic/management discussion.", "model": "claude-haiku-5-5"}
++{"work_id": 3093013316, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Legal and public-policy analysis of privacy law conflicts during COVID-19; not computer-science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3012891874, "categories": ["privacy_preserving_computation", "database_systems"], "primary": "privacy_preserving_computation", "confidence": 0.5, "rationale": "Graph-based analysis of identity PII assets and identity theft risk; privacy/identity protection with graph data analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 3004671875, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Location privacy attack and inference on geosocial network location sharing mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 2998582260, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Mimic learning to share models instead of sensitive health data, a privacy-preserving ML approach for smart health.", "model": "claude-haiku-5-5"}
++{"work_id": 4415149235, "categories": ["information_retrieval", "ml_theory"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Classifies GPS and check-in data into store categories, a location-based recommendation/classification task with positive-unlabeled learning; mainly data-driven geographic classification, low confidence on category.", "model": "claude-haiku-5-5"}
++{"work_id": 3217252216, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Synthesises Gaussian mechanisms to minimise mutual information for database privacy, a privacy-preserving data release problem with semidefinite programming.", "model": "claude-haiku-5-5"}
++{"work_id": 4401845216, "categories": ["federated_learning", "cybersecurity", "computer_networks"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Systematic review of privacy and security of federated learning in UAV systems; UAVs are a networked/robotic application context.", "model": "claude-haiku-5-5"}
++{"work_id": 4403764267, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.5, "rationale": "Discusses healthcare data privacy protection and leakage in cloud platforms, a data-protection topic with a weakly technical contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7155107746, "categories": ["generative_ai", "federated_learning"], "primary": "generative_ai", "confidence": 0.4, "rationale": "Selects synthetic samples generated by SDG models for downstream utility and transmission efficiency, using optimal control; primarily about synthetic data from generative models.", "model": "claude-haiku-5-5"}
++{"work_id": 4382045981, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Socio-business study of GDPR impact on business resilience; not a computer-science research contribution in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 4388951965, "categories": ["cybersecurity", "information_retrieval"], "primary": "cybersecurity", "confidence": 0.4, "rationale": "Social network analysis of the PGP web of trust for identifying influential users and key forgery concerns; security-related trust network analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4386204720, "categories": ["software_engineering", "cybersecurity"], "primary": "software_engineering", "confidence": 0.55, "rationale": "Source-code analysis with data-flow graphs to redact sensitive information in diagnostic logs, a software diagnosis and privacy tool.", "model": "claude-haiku-5-5"}
++{"work_id": 7165783687, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Survey-based study of consumer responses to GenAI-personalised advertising, privacy calculus and trust; a user-behaviour study rather than a technical contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7123336800, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Improves the exponential mechanism for differential privacy with priors and output constraints, a core DP mechanism design contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 2904199664, "categories": ["cybersecurity", "privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Privacy-preserving photo sharing in social networks via anonymization and trust; data-protection mechanism for shared content.", "model": "claude-haiku-5-5"}
++{"work_id": 4416037029, "categories": ["llm", "federated_learning"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Personalized federated fine-tuning of LoRA-based LLMs with hierarchical Bayesian aggregation.", "model": "claude-haiku-5-5"}
++{"work_id": 2952484830, "categories": ["privacy_preserving_computation", "database_systems"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "(k,p)-anonymity for sanitizing transactional databases to hide sensitive information.", "model": "claude-haiku-5-5"}
++{"work_id": 3204605723, "categories": ["privacy_preserving_computation", "cybersecurity"], "primary": "privacy_preserving_computation", "confidence": 0.45, "rationale": "Agent-based simulation of pandemic response using privacy-preserving encrypted contact chains on wearables; primarily epidemiological modelling with privacy framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4405979621, "categories": ["cybersecurity", "computer_vision"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Cancelable biometric template protection for privacy and security of biometric systems; biometric image processing.", "model": "claude-haiku-5-5"}
++{"work_id": 4385248498, "categories": ["hci", "privacy_preserving_computation"], "primary": "hci", "confidence": 0.7, "rationale": "Consumer resistance to intelligent voice assistants studied via experiments on privacy concerns and perceived creepiness.", "model": "claude-haiku-5-5"}
++{"work_id": 4412164648, "categories": ["federated_learning", "privacy_preserving_computation"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Differentially private client selection in federated learning balancing loss and privacy budget.", "model": "claude-haiku-5-5"}
++{"work_id": 2786217712, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.4, "rationale": "Historical and legislative review of encryption use and digital privacy policy; not a technical contribution, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7118160301, "categories": ["federated_learning", "ai_safety"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Privacy-preserving federated evaluation framework with statistical fairness testing across clients.", "model": "claude-haiku-5-5"}
++{"work_id": 4386192152, "categories": ["ai_safety"], "primary": null, "confidence": 0.1, "rationale": "Rule-based expert system for chicken disease diagnosis; does not clearly fit any ontology category, so no assignment given conservatively.", "model": "claude-haiku-5-5"}
++{"work_id": 4405359480, "categories": ["cybersecurity", "blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain-based access control and authentication for IoT data sharing via smart contracts; security framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3011050248, "categories": ["edge_computing", "ml_theory"], "primary": "edge_computing", "confidence": 0.35, "rationale": "ML-based co-location detection of IoT devices with cheap sensors; IoT sensing for logistics, weakly fits categories.", "model": "claude-haiku-5-5"}
++{"work_id": 2900068116, "categories": ["computer_networks", "reinforcement_learning"], "primary": "computer_networks", "confidence": 0.4, "rationale": "Participant selection in mobile crowdsensing with energy efficiency; MDP formulation is an optimization tool, mainly networking/sensing systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3090747742, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Manufacturing reliability modeling with sensor data; industrial engineering rather than a CS ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4214663994, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.3, "rationale": "Wearable sensor data fusion for medical AIoT; retracted notice, weak fit to health informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 3163035554, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.65, "rationale": "Task offloading and scheduling in wireless-powered mobile edge computing using an artificial fish swarm algorithm.", "model": "claude-haiku-5-5"}
++{"work_id": 7117133120, "categories": ["reinforcement_learning", "federated_learning", "edge_computing"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Federated transfer PPO for home energy management; RL method with federated learning component.", "model": "claude-haiku-5-5"}
++{"work_id": 3211844447, "categories": ["edge_computing", "cloud_computing"], "primary": "edge_computing", "confidence": 0.45, "rationale": "Software-defined smart factory fire protection spanning cloud/fog/edge; distributed infrastructure focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4410068037, "categories": ["cybersecurity", "hci"], "primary": "cybersecurity", "confidence": 0.4, "rationale": "Survey study of IoT security awareness among students; human factors of security and education.", "model": "claude-haiku-5-5"}
++{"work_id": 4413558883, "categories": ["applied_cryptography", "edge_computing"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Lightweight CP-ABE encryption for IoT with computation offloading to cooperative nodes.", "model": "claude-haiku-5-5"}
++{"work_id": 4383427890, "categories": ["computer_networks", "computer_vision"], "primary": "computer_networks", "confidence": 0.4, "rationale": "IoT mmWave positioning with AoA estimation and 3D localization; signal-processing/positioning framed as networking, with angle estimation from array signals.", "model": "claude-haiku-5-5"}
++{"work_id": 7126134868, "categories": ["cybersecurity", "reinforcement_learning", "ai_agents"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Malware defense in social IoT using multi-agent deep RL with Markov games and SHAP explanations.", "model": "claude-haiku-5-5"}
++{"work_id": 4389319530, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Conference preface on computational modeling and IoT; not a specific research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3007216950, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.3, "rationale": "Persian overview of Internet of Things as a networked sensor system; general introductory content.", "model": "claude-haiku-5-5"}
++{"work_id": 3025242379, "categories": ["information_retrieval", "computer_networks"], "primary": "computer_networks", "confidence": 0.35, "rationale": "Online incentive mechanism (reverse auction) for mobile crowdsensing worker selection; mechanism design for sensing systems, closer to networks than to other categories.", "model": "claude-haiku-5-5"}
++{"work_id": 4385153790, "categories": ["edge_computing", "computer_architecture"], "primary": "edge_computing", "confidence": 0.3, "rationale": "Hardware/software counting method for solar insecticidal lamps using machine learning, with lightweight edge computing motivation; applied consumer electronics.", "model": "claude-haiku-5-5"}
++{"work_id": 4281393175, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Computation offloading and resource allocation in mobile edge computing to minimize device energy.", "model": "claude-haiku-5-5"}
++{"work_id": 4312555758, "categories": ["edge_computing", "computational_biology"], "primary": "edge_computing", "confidence": 0.3, "rationale": "IoT wearable monitoring platform for prehabilitation; primarily an IoT system in healthcare, weakly edge/health informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 3096106062, "categories": ["edge_computing", "deep_learning_architectures"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Tiny machine learning on embedded and IoT devices with pruning and approximation for on-device incremental learning.", "model": "claude-haiku-5-5"}
++{"work_id": 3199685485, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Low-power sensor readout circuit with switching DC-DC regulator for IoT devices; hardware design.", "model": "claude-haiku-5-5"}
++{"work_id": 3216809321, "categories": ["cybersecurity", "applied_cryptography", "distributed_computing"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Securing cloud-IoT data via encryption (RC6) in private cloud; security of IoT systems is the main focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4401883192, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Book chapter on ML-driven discovery in IoT; overview without a clear category fit, weakly ML-related.", "model": "claude-haiku-5-5"}
++{"work_id": 4390284844, "categories": ["edge_computing", "federated_learning", "blockchain"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Resource allocation for edge servers running blockchain-based federated learning and MEC offloading.", "model": "claude-haiku-5-5"}
++{"work_id": 4400095687, "categories": ["cloud_computing", "edge_computing", "blockchain"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Architecture integrating cloud, fog, edge and blockchain for SDG data governance.", "model": "claude-haiku-5-5"}
++{"work_id": 3176753072, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Survey of IoT security requirements, vulnerabilities and encryption mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 3184417063, "categories": ["speech_audio", "ml_theory"], "primary": null, "confidence": 0.25, "rationale": "Contactless non-intrusive load monitoring via sensing and modality distillation; sensing signal processing, weakly fits categories.", "model": "claude-haiku-5-5"}
++{"work_id": 4390905002, "categories": ["edge_computing", "deep_learning_architectures"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Model deployment and task offloading for multi-task inference in vehicular edge computing.", "model": "claude-haiku-5-5"}
++{"work_id": 7156021111, "categories": ["edge_computing", "ml_theory"], "primary": "edge_computing", "confidence": 0.4, "rationale": "IoT hardware gateway design with soil moisture prediction; edge-IoT platform focus with ML modeling.", "model": "claude-haiku-5-5"}
++{"work_id": 4399525972, "categories": ["edge_computing", "deep_learning_architectures"], "primary": "edge_computing", "confidence": 0.75, "rationale": "End-edge collaborative inference and partitioning of convolutional fuzzy neural networks for IoT.", "model": "claude-haiku-5-5"}
++{"work_id": 2886809270, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Lightweight stream-cipher-based cryptographic scheme with dynamic key for IoT communication.", "model": "claude-haiku-5-5"}
++{"work_id": 4289520433, "categories": ["edge_computing", "reinforcement_learning"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Vehicular edge computing task offloading with end-edge-cloud collaboration, using DDPG RL for offloading policy.", "model": "claude-haiku-5-5"}
++{"work_id": 4367394805, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.4, "rationale": "IoT smart healthcare architecture; IoT platform as computing substrate, weakly fits edge; mostly a general IoT system overview.", "model": "claude-haiku-5-5"}
++{"work_id": 7164347165, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Attention-guided federated learning framework for energy-efficient green computing on edge devices.", "model": "claude-haiku-5-5"}
++{"work_id": 4378364725, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "IoT security system integrated with blockchain; literature review of IoT and blockchain protocols.", "model": "claude-haiku-5-5"}
++{"work_id": 4411194662, "categories": ["ai_safety", "ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Distributionally robust stochastic MPC for cyber-physical systems under disturbances and DoS attacks; control theory mostly, security aspect secondary, unclear fit to ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 4411000650, "categories": ["edge_computing", "computer_vision"], "primary": "edge_computing", "confidence": 0.35, "rationale": "IoT sensor data with deep learning for early warning in infrastructure; edge and scalable deep learning focus, sensor data analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 3047548318, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Survey of fog computing applied in IoT environments.", "model": "claude-haiku-5-5"}
++{"work_id": 4281719112, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Electronic nose hardware prototype for gas sensing; embedded sensor hardware, weakly fits any CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4410347576, "categories": ["edge_computing", "reinforcement_learning"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Deep Q-learning based task offloading and resource management in fog computing.", "model": "claude-haiku-5-5"}
++{"work_id": 3097229291, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Deep RL with a Fisher information reward for adaptive sensing policies on IoT devices.", "model": "claude-haiku-5-5"}
++{"work_id": 4410459414, "categories": ["ml_theory", "database_systems"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Literature review of predictive maintenance in Industry 4.0 with ML integration and benchmark dataset challenges; ontology has no clear industrial-maintenance category, so classification is weak and low-confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 2902217519, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.9, "rationale": "Security evaluation of smart home platforms revealing vulnerabilities and privilege escalation.", "model": "claude-haiku-5-5"}
++{"work_id": 4416583070, "categories": ["edge_computing", "privacy_preserving_computation", "federated_learning"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Edge-based real-time patient monitoring with privacy-preserving encryption and federated learning for healthcare.", "model": "claude-haiku-5-5"}
++{"work_id": 2918201812, "categories": ["edge_computing", "information_retrieval"], "primary": "edge_computing", "confidence": 0.5, "rationale": "User-centric location prediction applied to mobile edge service migration; location prediction is ML-driven but application targets MEC.", "model": "claude-haiku-5-5"}
++{"work_id": 4294691681, "categories": ["edge_computing", "deep_learning_architectures", "computer_architecture"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Lightweight STFT CNN for ECG arrhythmia detection deployed on edge devices with quantization and pruning; edge deployment is central, with architecture contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4415397548, "categories": ["generative_ai", "computer_architecture"], "primary": "generative_ai", "confidence": 0.3, "rationale": "Systematic review of Generative AI and IoT in precision agriculture; the focus spans IoT and agriculture, so the classification is tentative.", "model": "claude-haiku-5-5"}
++{"work_id": 4409047392, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Authentication and encryption frameworks (RSA, AES, ECC) for IoT data security.", "model": "claude-haiku-5-5"}
++{"work_id": 4379398101, "categories": ["cloud_computing", "edge_computing"], "primary": "edge_computing", "confidence": 0.7, "rationale": "PSO-based workload allocation between fog devices and cloud data centers to reduce power and delay.", "model": "claude-haiku-5-5"}
++{"work_id": 7123925437, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Task allocation in vehicular edge computing using Lion Optimization Algorithm.", "model": "claude-haiku-5-5"}
++{"work_id": 4413824611, "categories": ["edge_computing", "cybersecurity"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Fog-based secure healthcare data management over 5G, focused on fog computing and data security.", "model": "claude-haiku-5-5"}
++{"work_id": 4377038089, "categories": ["computer_vision", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Survey of face morph attack detection, a security problem applied to face images.", "model": "claude-haiku-5-5"}
++{"work_id": 4362518865, "categories": ["cybersecurity", "reinforcement_learning"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Black-box adversarial attack on SQL injection detectors; security focus with RL-based generation.", "model": "claude-haiku-5-5"}
++{"work_id": 4401110449, "categories": ["cybersecurity", "generative_ai"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Phishing detection with an improved GAN classifier; security application.", "model": "claude-haiku-5-5"}
++{"work_id": 3090172217, "categories": ["cybersecurity", "privacy_preserving_computation"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Cloud data confidentiality via hybrid attribute-based re-encryption; encryption for data protection in cloud.", "model": "claude-haiku-5-5"}
++{"work_id": 7171536369, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.9, "rationale": "Safety alignment of LLMs against neuron-pruning attacks.", "model": "claude-haiku-5-5"}
++{"work_id": 2952116314, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Empirical study of DDoS attack impact on cryptocurrency exchange trading.", "model": "claude-haiku-5-5"}
++{"work_id": 3197514244, "categories": ["cybersecurity", "software_engineering"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Honeypot design for microservices architectures to collect attack data.", "model": "claude-haiku-5-5"}
++{"work_id": 7219206452, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Analysis of cyber crimes and legal control in India; social-legal framing, borderline CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3167654854, "categories": ["cybersecurity", "information_retrieval"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Machine learning model to detect fraudulent posts in C2C marketplaces; fraud detection framed as security.", "model": "claude-haiku-5-5"}
++{"work_id": 3044959568, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Survey-based study of employee information security policy compliance in government.", "model": "claude-haiku-5-5"}
++{"work_id": 2944025354, "categories": ["applied_cryptography", "distributed_computing", "cloud_computing"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Key management and encryption scheme for cloud data using ECC, Diffie-Hellman and OTP; primarily cryptographic protocol design.", "model": "claude-haiku-5-5"}
++{"work_id": 4390504465, "categories": ["cybersecurity", "deep_learning_architectures", "speech_audio"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "EEG motor imagery biometric authentication with hybrid CNN models; biometric security is the framing, but signal-based classification is also present. Brain signal is not audio, so speech_audio is a weak fit and is dropped.", "model": "claude-haiku-5-5"}
++{"work_id": 3031484172, "categories": ["hci"], "primary": "hci", "confidence": 0.8, "rationale": "Mixed-methods user study of smart home device sharing and access control needs.", "model": "claude-haiku-5-5"}
++{"work_id": 7170084806, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Behavior-based continuous user authentication using autoencoder and PSO for security.", "model": "claude-haiku-5-5"}
++{"work_id": 4200300237, "categories": [], "primary": null, "confidence": 0.8, "rationale": "News report on a data breach; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7154943096, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Machine learning phishing URL detection using lexical and structural features.", "model": "claude-haiku-5-5"}
++{"work_id": 4412549228, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Image encryption using chaotic cryptography combined with steganography.", "model": "claude-haiku-5-5"}
++{"work_id": 4392971839, "categories": ["llm", "ai_safety", "reinforcement_learning"], "primary": "llm", "confidence": 0.7, "rationale": "Watermarking LLM outputs by tuning LLM weights with RL; contributions concern LLM text provenance, which also fits generative and safety framings.", "model": "claude-haiku-5-5"}
++{"work_id": 7167720828, "categories": ["privacy_preserving_computation", "cybersecurity", "cloud_computing"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Security framework for privacy and data protection in outsourced data, combining key generation, masking and anonymisation.", "model": "claude-haiku-5-5"}
++{"work_id": 4416409576, "categories": ["generative_ai", "multimodal_ai", "cybersecurity"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Multimodal deepfake and fake news detection; AI-generated content detection with multimodal fusion.", "model": "claude-haiku-5-5"}
++{"work_id": 7214148734, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Black-box adversarial evasion attack on MalConv malware detector; malware and detection evasion is cybersecurity, with adversarial robustness of ML as secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4285320914, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Digital forensics tools for information security of organisations.", "model": "claude-haiku-5-5"}
++{"work_id": 7171969272, "categories": ["generative_ai", "cybersecurity", "computer_vision"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Deepfake image detection with a VGG16-CNN; AI-generated content detection is generative_ai, with image classification and security framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3083517685, "categories": ["cybersecurity", "hci"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Digital identity services for smartphone users addressing fraud resistance and privacy; abstract is minimal.", "model": "claude-haiku-5-5"}
++{"work_id": 4298144048, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Comparative evaluation of web application security scanners for vulnerability detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4310682878, "categories": ["cybersecurity", "information_retrieval"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Detection of malware and search rank fraud on Google Play; fraud and malware detection framed as security.", "model": "claude-haiku-5-5"}
++{"work_id": 7011534987, "categories": ["cybersecurity", "database_systems"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Sequential pattern mining in phone recordings for digital forensics using answer set programming; forensic application with data mining.", "model": "claude-haiku-5-5"}
++{"work_id": 7131118580, "categories": ["cybersecurity", "ml_theory"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Uncertainty-aware deep kernel ensemble for web attack detection; security application with uncertainty estimation.", "model": "claude-haiku-5-5"}
++{"work_id": 7200539267, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Software development for secure transmission and reception of information; cryptographic information protection.", "model": "claude-haiku-5-5"}
++{"work_id": 4385221456, "categories": ["nlp", "cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Hybrid CNN-BiLSTM-attention with FastText for Korean voice phishing detection; security application using text classification.", "model": "claude-haiku-5-5"}
++{"work_id": 4393303862, "categories": ["cybersecurity", "software_engineering"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Threat modeling of cyber-physical systems and secure software engineering practices; security is the core topic with an SE empirical component.", "model": "claude-haiku-5-5"}
++{"work_id": 4239390297, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Malicious URL detection is a security task; the deep learning angle is the method.", "model": "claude-haiku-5-5"}
++{"work_id": 2990181665, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Protective security overview; general security rather than a clear CS research contribution, so conservative single label.", "model": "claude-haiku-5-5"}
++{"work_id": 7154473294, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "XAI-driven threat analysis in SIEM systems for cybersecurity operations.", "model": "claude-haiku-5-5"}
++{"work_id": 7215892371, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Comparative analysis of Android malware detection methods, including neural approaches.", "model": "claude-haiku-5-5"}
++{"work_id": 2816623457, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "DoS intrusion detection for cloud computing environments.", "model": "claude-haiku-5-5"}
++{"work_id": 4380988124, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Cyber security architecture analysis for smart buildings and IoT systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3162129019, "categories": ["cybersecurity", "computer_architecture"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Exploiting radiation-induced soft errors for attacks on classical and quantum computers; hardware fault exploitation.", "model": "claude-haiku-5-5"}
++{"work_id": 3112992601, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Survey-based study of COVID-19 impact on corporate cybersecurity risk management.", "model": "claude-haiku-5-5"}
++{"work_id": 7001137095, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Remote access VPN configuration and security on a router, spanning security and networking.", "model": "claude-haiku-5-5"}
++{"work_id": 4289655646, "categories": ["cybersecurity", "applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Post-quantum RLWE-based certificateless key exchange protocol with security proof; cryptographic protocol design.", "model": "claude-haiku-5-5"}
++{"work_id": 3089767340, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Engaging users with feedback to choose stronger passwords; user-centred interaction study.", "model": "claude-haiku-5-5"}
++{"work_id": 4411782417, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Malware classification framework for cloud security using feature selection and imbalance handling.", "model": "claude-haiku-5-5"}
++{"work_id": 7155614231, "categories": ["cybersecurity", "hci"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Survey-based study of measures preventing social engineering attacks on mobile customers.", "model": "claude-haiku-5-5"}
++{"work_id": 2995203890, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.55, "rationale": "DWT-based image watermarking; image processing with security aspect, closest to computer vision/multimedia.", "model": "claude-haiku-5-5"}
++{"work_id": 4409103867, "categories": ["nlp", "cybersecurity"], "primary": "nlp", "confidence": 0.6, "rationale": "LSTM-based text classification for email spam detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4393105534, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Interactive educational media design on sex education for adolescents; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4409248207, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Zero-trust continuous authentication for IoT using identity-based encryption.", "model": "claude-haiku-5-5"}
++{"work_id": 4403677606, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Risk analysis methodology for GenAI systems addressing prompt injection threats to LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 3041436169, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "RNN classification of power grid contingencies and cyber-attacks for intrusion detection.", "model": "claude-haiku-5-5"}
++{"work_id": 7160546262, "categories": ["software_engineering", "ai_agents", "ml_theory"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Software remodularization framed as multi-agent negotiation; core is software module clustering (SE) using agent consensus protocol (ai_agents) with formal properties.", "model": "claude-haiku-5-5"}
++{"work_id": 4311440938, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.45, "rationale": "Deployment/hosting architecture for MATLAB web apps; closest to software systems engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 4409591325, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Agile methodology for large-scale IT transformation, project management and remote teams.", "model": "claude-haiku-5-5"}
++{"work_id": 7128328612, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.35, "rationale": "Implementation of an online store on a CMS platform; web development thesis, weakly SE.", "model": "claude-haiku-5-5"}
++{"work_id": 6986525772, "categories": ["formal_verification", "blockchain", "programming_languages"], "primary": "blockchain", "confidence": 0.7, "rationale": "Pre-deployment analysis of smart contracts, data races and static analysis for Ethereum; blockchain security and verification.", "model": "claude-haiku-5-5"}
++{"work_id": 4200040010, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Book chapter on Python unittest test automation, a testing topic.", "model": "claude-haiku-5-5"}
++{"work_id": 3157577106, "categories": ["software_engineering", "ml_theory"], "primary": "software_engineering", "confidence": 0.65, "rationale": "Software effort estimation with COCOMO optimized by PSO hybrid with SVM/LR/RF.", "model": "claude-haiku-5-5"}
++{"work_id": 7118011917, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Book chapter on programming functions; introductory programming, weakly SE.", "model": "claude-haiku-5-5"}
++{"work_id": 2972326490, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Design of an Android information application with web service using waterfall SDLC.", "model": "claude-haiku-5-5"}
++{"work_id": 2932905330, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.85, "rationale": "Tool support for test smell summaries improving test suite quality awareness.", "model": "claude-haiku-5-5"}
++{"work_id": 3008788199, "categories": ["nlp", "ml_theory", "software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Ontology-based ML classification of security requirements in requirements engineering; text classification with requirements focus, so software engineering primary with NLP-style classification.", "model": "claude-haiku-5-5"}
++{"work_id": 4387781545, "categories": ["software_engineering", "hci"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Structured editors and visual replacements for source code are developer tooling for programming environments.", "model": "claude-haiku-5-5"}
++{"work_id": 4401907809, "categories": ["software_engineering", "ai_agents", "reinforcement_learning"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Testing autonomous driving systems with RL-driven adversarial agents; primarily software testing of ADS.", "model": "claude-haiku-5-5"}
++{"work_id": 4410538665, "categories": ["llm", "software_engineering"], "primary": "llm", "confidence": 0.7, "rationale": "Benchmark and evaluation of instruction-tuned code-generating LLMs for correctness and robustness.", "model": "claude-haiku-5-5"}
++{"work_id": 7219410480, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Project management methodologies for IT software development project.", "model": "claude-haiku-5-5"}
++{"work_id": 7166522127, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "White-label reuse approach for developing mobile e-commerce applications for MSMEs.", "model": "claude-haiku-5-5"}
++{"work_id": 3143264371, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Quality assurance framework for software process and product in very small entities per ISO standards.", "model": "claude-haiku-5-5"}
++{"work_id": 4392404200, "categories": ["software_engineering", "ml_theory"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Weighted ensemble learning for software project failure prediction; software project management focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4253254765, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Spreadsheet book chapter on Excel tables; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4310841548, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Automatic generation of REST API backend projects via templates; code generation tooling for application development.", "model": "claude-haiku-5-5"}
++{"work_id": 4386014351, "categories": ["information_retrieval"], "primary": null, "confidence": 0.2, "rationale": "Construction procurement digitalization framework; business/IT topic with weak CS fit, marginally related to supplier relationship systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4280590664, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Business transformation and agile management framework; not CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 4312839447, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Book chapter on building Android apps; tutorial content with no research contribution in an ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4206446312, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Decentralized multi-agent control system for assembly; agent architecture, with manufacturing/robotics context.", "model": "claude-haiku-5-5"}
++{"work_id": 4316021200, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Code smell identification and refactoring to improve mobile app performance.", "model": "claude-haiku-5-5"}
++{"work_id": 4388972704, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Composition of services across heterogeneous blockchains.", "model": "claude-haiku-5-5"}
++{"work_id": 7215194540, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Running Jasmine JavaScript tests via Maven build tool; software testing tooling.", "model": "claude-haiku-5-5"}
++{"work_id": 4293177572, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Web application for ISP front-office management; a business information system with weak research framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7203691417, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Evaluates lazy learning algorithms for software effort estimation; ML methodology applied to software engineering, could also fit software_engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 7135050868, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Defines modifiability scenario for quantum software maintainability from version histories.", "model": "claude-haiku-5-5"}
++{"work_id": 3044591241, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Development of an Android attendance application; an application-building system paper with no clear fit to a specific research category beyond software development.", "model": "claude-haiku-5-5"}
++{"work_id": 7217619139, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Design and development of an Android academic monitoring application using standard web and mobile frameworks.", "model": "claude-haiku-5-5"}
++{"work_id": 7162132995, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Antibody CDR design with structural and interface prediction is a protein structure design problem using geometric deep learning.", "model": "claude-haiku-5-5"}
++{"work_id": 4385395724, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Architecture design and implementation of an SME information system using SOA and microservices.", "model": "claude-haiku-5-5"}
++{"work_id": 2908635758, "categories": ["nlp"], "primary": null, "confidence": 0.1, "rationale": "Mathematics education paper on simulations of probability paradoxes; not clearly a computer-science research category, so no category is assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 7113083064, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Qualitative sociotechnical study of student and teacher perspectives on generative AI in programming education, centred on human factors.", "model": "claude-haiku-5-5"}
++{"work_id": 2981150652, "categories": ["hci", "multimodal_ai"], "primary": "hci", "confidence": 0.7, "rationale": "Browser-based prototyping environment for novices to build ML apps within webpages, an interactive systems and end-user tooling contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7081987054, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.9, "rationale": "Action research on establishing technical debt management in an IT company.", "model": "claude-haiku-5-5"}
++{"work_id": 4239998720, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Web-based futsal rental information system built with CodeIgniter.", "model": "claude-haiku-5-5"}
++{"work_id": 3026019741, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Web application design for dormitory room data management using the waterfall method.", "model": "claude-haiku-5-5"}
++{"work_id": 7116117897, "categories": ["llm", "software_engineering"], "primary": "llm", "confidence": 0.8, "rationale": "Evaluates LLM code reasoning abilities via benchmarks; also code-analysis dataset construction tied to software engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 7163811937, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Statistical methodology on bootstrap inference and resampling with imbalance; principled ML/statistics methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 2859484040, "categories": ["software_engineering", "ai_safety"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Metamorphic testing to find implementation bugs in ML applications is software testing research.", "model": "claude-haiku-5-5"}
++{"work_id": 7154701096, "categories": ["software_engineering", "hci"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Web application system for team collaboration and workflow automation; system design and usability evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 7117153023, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.35, "rationale": "Development of a monitoring web tool with microservices and usability testing; software development thesis.", "model": "claude-haiku-5-5"}
++{"work_id": 2931497058, "categories": ["hci"], "primary": "hci", "confidence": 0.25, "rationale": "Android tour guide application built with Google Maps API; application development with black-box testing, weakly fits any category.", "model": "claude-haiku-5-5"}
++{"work_id": 4205450578, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.4, "rationale": "Book chapter on generic functions and types, a programming language feature.", "model": "claude-haiku-5-5"}
++{"work_id": 4398220511, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Web-based information system for teacher data management; the abstract is only a reference list.", "model": "claude-haiku-5-5"}
++{"work_id": 7155216780, "categories": ["software_engineering", "ai_safety"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Frames machine unlearning testing as a software QA problem; also concerns privacy and model safety.", "model": "claude-haiku-5-5"}
++{"work_id": 4405173371, "categories": ["software_engineering", "information_retrieval"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Information-theoretic evaluation of unsupervised software traceability via IR techniques.", "model": "claude-haiku-5-5"}
++{"work_id": 4383108538, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.8, "rationale": "Task-aware grasp estimation and placement for robot manipulation in cluttered scenes; uses perspective affordance maps from vision.", "model": "claude-haiku-5-5"}
++{"work_id": 4403919332, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.8, "rationale": "Legged robot semantic mapping for search and rescue with a visual recognition model integrated into navigation.", "model": "claude-haiku-5-5"}
++{"work_id": 4410526140, "categories": ["robotics", "computer_vision", "deep_learning_architectures"], "primary": "robotics", "confidence": 0.6, "rationale": "Educational SCARA robot with vision-based control and CNN/autoencoder fault classification; robotics is the core platform.", "model": "claude-haiku-5-5"}
++{"work_id": 3182010183, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Tethered UAV relay design for control and non-payload communications over wireless links.", "model": "claude-haiku-5-5"}
++{"work_id": 4388100051, "categories": ["deep_learning_architectures", "robotics"], "primary": "robotics", "confidence": 0.6, "rationale": "Graph neural network model for vehicle path forecasting in automated driving; autonomous vehicle context, GNN method.", "model": "claude-haiku-5-5"}
++{"work_id": 4394659963, "categories": ["robotics", "ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Chapter on trust, safety, assurance and regulation of robotics and AI systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4385413558, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.6, "rationale": "Backstepping-based interception control for an underactuated USV with adaptive estimation and experiments.", "model": "claude-haiku-5-5"}
++{"work_id": 4315632170, "categories": ["hci", "ai_safety"], "primary": "hci", "confidence": 0.4, "rationale": "Ethics of social robots in eldercare from a Confucian perspective; philosophical and human-robot interaction focus, not clearly CS core.", "model": "claude-haiku-5-5"}
++{"work_id": 3173643949, "categories": ["computer_vision", "algorithms"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Graph optimization via ICM for visual place recognition; fast algorithm applied to image retrieval for localization.", "model": "claude-haiku-5-5"}
++{"work_id": 4404604190, "categories": ["robotics", "algorithms"], "primary": "robotics", "confidence": 0.6, "rationale": "Position analysis of mechanisms via multi-affine formulations and interval methods for kinematic closure problems in robotics.", "model": "claude-haiku-5-5"}
++{"work_id": 4413925424, "categories": ["robotics", "ai_agents", "ml_theory"], "primary": "robotics", "confidence": 0.7, "rationale": "Decentralized safe multi-agent control of robots with control barrier functions; robotics primary, multi-agent control secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 2909162014, "categories": ["robotics", "deep_learning_architectures"], "primary": "robotics", "confidence": 0.7, "rationale": "Industrial robot threading task failure prediction using CNN classifier on force-torque signals.", "model": "claude-haiku-5-5"}
++{"work_id": 4415135180, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Scenario-based certification of automated vehicles using UAV field data; autonomous vehicles fall under robotics.", "model": "claude-haiku-5-5"}
++{"work_id": 4313471462, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.6, "rationale": "Path tracking control for autonomous vehicles with PSO-BP adaptive MPC weights.", "model": "claude-haiku-5-5"}
++{"work_id": 4319034475, "categories": ["robotics"], "primary": "robotics", "confidence": 0.6, "rationale": "Fault-tolerant control of underwater salvage robot with event triggering.", "model": "claude-haiku-5-5"}
++{"work_id": 4416749351, "categories": ["ai_agents", "robotics"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Multi-agent pathfinding improving conflict-based search via conflict area prediction.", "model": "claude-haiku-5-5"}
++{"work_id": 7214885326, "categories": ["robotics", "multimodal_ai"], "primary": "robotics", "confidence": 0.5, "rationale": "Multi-UAV coordination with RGB and event vision; aerial robot communication mechanism.", "model": "claude-haiku-5-5"}
++{"work_id": 4382935921, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Control theory for nonlinear networked systems with delays, contraction theory; closest category is ml_theory-free control, weakly fits none well.", "model": "claude-haiku-5-5"}
++{"work_id": 2909059777, "categories": ["robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "Fuzzy cooperative behaviors for multi-robot flocking and formation compared to Boids.", "model": "claude-haiku-5-5"}
++{"work_id": 7161591331, "categories": ["robotics", "reinforcement_learning", "hci"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "RL with human preference guidance for robot manipulation; RL primary with robotics application.", "model": "claude-haiku-5-5"}
++{"work_id": 4297857948, "categories": ["computer_vision", "robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "Recovers 3D cable models from single RGB images to support robotic cable untangling; combines 2D/3D neural recognition with simulation for manipulation.", "model": "claude-haiku-5-5"}
++{"work_id": 3034448848, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Fault detection and isolation for UAS air data systems; closest fit is avionics/sensor fault estimation, which is not cleanly in any ontology category, so only a low-confidence assignment is made.", "model": "claude-haiku-5-5"}
++{"work_id": 4394880330, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Estimates aerosol optical depth from ground air-quality data for environmental/atmospheric science; a computational estimation method applied to earth science, not clearly CS core.", "model": "claude-haiku-5-5"}
++{"work_id": 4391947971, "categories": ["robotics", "hci"], "primary": "robotics", "confidence": 0.7, "rationale": "Multi-modal robotic skin with proximity and tactile sensors for human-robot interaction and collaboration in industrial settings.", "model": "claude-haiku-5-5"}
++{"work_id": 4307184398, "categories": ["hci"], "primary": null, "confidence": 0.3, "rationale": "Survey of preservice teachers' opinions on computational thinking in preschool; education study with weak CS-category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4313227226, "categories": ["ai_agents"], "primary": null, "confidence": 0.35, "rationale": "Multi-UAV coordinated trajectory planning via a metaheuristic swarm-style algorithm; best fit is autonomous multi-agent planning, though it is optimisation-centred.", "model": "claude-haiku-5-5"}
++{"work_id": 4382059457, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Experiments on customer perceptions of frontline service robots versus employees; human-AI/robot interaction in a service-marketing framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4410227174, "categories": ["robotics"], "primary": null, "confidence": 0.35, "rationale": "EKF with adaptive digital filter for non-stationary signal filtering, motivated by robotic state observers; signal processing with robotics framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4403277362, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.6, "rationale": "Learning from demonstration of movement primitives with mixture-of-experts networks, validated on real robot tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 4281717654, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Semantic grid map-assisted AMCL localization for mobile robots in large similar environments.", "model": "claude-haiku-5-5"}
++{"work_id": 2903260265, "categories": ["ai_agents", "robotics"], "primary": "robotics", "confidence": 0.6, "rationale": "Behavior planning for autonomous vehicles under uncertainty using MCTS and QMDP; decision-making for autonomous driving with some agent-planning flavor.", "model": "claude-haiku-5-5"}
++{"work_id": 4210424983, "categories": ["robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "Path planning for an unmanned ship using sensor fusion and firefly optimisation; a robotic system for navigation.", "model": "claude-haiku-5-5"}
++{"work_id": 4283759396, "categories": ["robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "Miniature autonomous vehicles for testing autonomous driving methods; physical robotic vehicles.", "model": "claude-haiku-5-5"}
++{"work_id": 4412030501, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Bimanual robot manipulation primitives for deformable food objects.", "model": "claude-haiku-5-5"}
++{"work_id": 2886313903, "categories": ["robotics"], "primary": "robotics", "confidence": 0.75, "rationale": "Control of an active SLIP legged running model, a legged locomotion robotics control problem.", "model": "claude-haiku-5-5"}
++{"work_id": 3011793074, "categories": ["robotics"], "primary": "robotics", "confidence": 0.75, "rationale": "Model predictive path tracking control for autonomous vehicles, validated with field experiments.", "model": "claude-haiku-5-5"}
++{"work_id": 4367860403, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Mobile quad-arm wheeled-legged robot design and locomotion.", "model": "claude-haiku-5-5"}
++{"work_id": 4403447662, "categories": ["robotics"], "primary": "robotics", "confidence": 0.75, "rationale": "In-hand dexterous manipulation of unknown objects with a robotic hand using tactile sensing.", "model": "claude-haiku-5-5"}
++{"work_id": 4381429872, "categories": ["robotics"], "primary": "robotics", "confidence": 0.65, "rationale": "UAV obstacle avoidance and patrol path planning with ultrasonic and multi-sensor fusion.", "model": "claude-haiku-5-5"}
++{"work_id": 4391695417, "categories": ["robotics"], "primary": "robotics", "confidence": 0.65, "rationale": "Localization and slip compensation for autonomous robot navigation.", "model": "claude-haiku-5-5"}
++{"work_id": 3006755905, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.8, "rationale": "Four-wheel mobile robot with gesture-controlled arm and object detection for pick and place; physical robot system with some vision component.", "model": "claude-haiku-5-5"}
++{"work_id": 7160616345, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Offline-to-online RL with action chunking and adaptive critic selection; applied to robot manipulation benchmarks.", "model": "claude-haiku-5-5"}
++{"work_id": 4392494944, "categories": ["robotics", "hci"], "primary": "robotics", "confidence": 0.7, "rationale": "Habilitation summary on robot learning and interaction for human-humanoid collaboration.", "model": "claude-haiku-5-5"}
++{"work_id": 2788120109, "categories": ["robotics", "hci"], "primary": "robotics", "confidence": 0.75, "rationale": "Adjustable autonomy teleoperation architecture for robots with usability study.", "model": "claude-haiku-5-5"}
++{"work_id": 4415293255, "categories": ["robotics"], "primary": "robotics", "confidence": 0.9, "rationale": "LiDAR-inertial-kinematic state estimation for humanoid robots.", "model": "claude-haiku-5-5"}
++{"work_id": 7160828610, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Plant modeling and system identification for an indoor autonomous robot blimp pointing control.", "model": "claude-haiku-5-5"}
++{"work_id": 3126252149, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.6, "rationale": "Filter-based ego-vehicle motion estimation for autonomous race car; state estimation with sensor fusion.", "model": "claude-haiku-5-5"}
++{"work_id": 4403010033, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Multi-robot path planning fusing DWA and A* in unknown indoor space.", "model": "claude-haiku-5-5"}
++{"work_id": 3130719769, "categories": ["robotics", "computer_architecture"], "primary": "robotics", "confidence": 0.4, "rationale": "Kinematic chain estimation of an excavator using bacterial programming and INS sensors; closest to robotics.", "model": "claude-haiku-5-5"}
++{"work_id": 7135139717, "categories": ["robotics", "reinforcement_learning"], "primary": "robotics", "confidence": 0.75, "rationale": "Real2Sim2Real transformer policy for bimanual robot manipulation and human-robot handover tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 7158557735, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.8, "rationale": "Probability/convex-order sub-Gaussian comparison result; statistical theory of concentration, no clear ML application beyond theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4410294818, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Regularizer for deep regression networks addressing imbalanced regression; methodological neural network training contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7215048402, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied comparison of WEKA resampling for classification on a breast cancer dataset; mainly applied ML/medical data analysis without a clear ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4391093173, "categories": ["federated_learning", "privacy_preserving_computation"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Analysis of differential privacy in federated learning via output perturbation and DP-SGD.", "model": "claude-haiku-5-5"}
++{"work_id": 7219508654, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied random forest classification of medical students' scores; educational data mining with no clear ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 7133993797, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Deep credal clustering using Dempster-Shafer theory for uncertainty; clustering method with theoretical derivations, borderline with deep learning.", "model": "claude-haiku-5-5"}
++{"work_id": 3179767201, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Reject-option classifiers with calibration, excess risk and generalization bounds; theory-focused with deep architectures.", "model": "claude-haiku-5-5"}
++{"work_id": 7163158107, "categories": ["ml_theory", "ai_safety"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Dropout prediction with interpretability and fairness testing; applied ML emphasizing fairness, not clearly in a single category.", "model": "claude-haiku-5-5"}
++{"work_id": 3155002205, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Numerical methods for partial SVD and rank estimation via Krylov subspaces; matrix algorithm design, possibly scientific computing.", "model": "claude-haiku-5-5"}
++{"work_id": 7160115409, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Bayesian conformal prediction with theoretical guarantees and regret bounds under distribution shift.", "model": "claude-haiku-5-5"}
++{"work_id": 3126933325, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Optimistic Q-learning agent with regret analysis for general agent-environment interface; RL theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4327500503, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Active learning system for imbalanced streams with class discovery; primarily ML methodology, could also be ai_agents or database-like but best fit is active learning under ml_theory.", "model": "claude-haiku-5-5"}
++{"work_id": 7171392614, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Industrial process control via ensemble predictive modeling and Bayesian optimization; applied ML, closest ontology fit is generic ML methodology with low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7214886155, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Biclustering combined with causal discovery under DAG constraints; causal inference methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4402510141, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Inductive inference operators (system W, rational closure) from belief bases; formal knowledge representation and nonmonotonic reasoning, closest to ML theory/logic.", "model": "claude-haiku-5-5"}
++{"work_id": 4388726461, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Active learning sampling strategy for quality inspection in manufacturing; ML methodology applied to industry.", "model": "claude-haiku-5-5"}
++{"work_id": 7170116302, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.25, "rationale": "MLP for predicting student performance with feature selection and optimizer comparison; applied ML, low-confidence fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4385985189, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Ensemble ML for mineral prospectivity mapping; geoscience application, weak fit to ontology, scientific_computing alternative.", "model": "claude-haiku-5-5"}
++{"work_id": 7151296083, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Deep anomaly detection on images via pretext tasks and fraud detection application; visual anomaly detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4296563198, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Bayesian framework for feature importance explanations with privacy and fairness tradeoffs; interpretability as formal problem.", "model": "claude-haiku-5-5"}
++{"work_id": 7143524691, "categories": ["ai_safety", "generative_ai"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Framework integrating responsible and explainable AI with generative AI; focus on bias, interpretability, governance places it mainly in AI safety, with generative AI as a secondary theme.", "model": "claude-haiku-5-5"}
++{"work_id": 4236967208, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Book on kernel methods, a statistical learning theory topic covering kernel trick and dual representations.", "model": "claude-haiku-5-5"}
++{"work_id": 4409046524, "categories": ["reinforcement_learning", "algorithms", "ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Clustering with bandit feedback concerns computational-statistical tradeoffs in online/bandit learning; abstract is empty so classification is uncertain.", "model": "claude-haiku-5-5"}
++{"work_id": 4408365506, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Music education model built from student behavioural data and logistic regression; not a computer-science research contribution in the ontology sense, so no category assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 4305012164, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Bayesian reversible jump MCMC for mixture models estimated from grouped income data; primarily statistical methodology with probabilistic models.", "model": "claude-haiku-5-5"}
++{"work_id": 4404916629, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Fuzzy clustering algorithms for relational data with regularization; clustering methodology best fit to probabilistic/statistical learning, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 2737706773, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Large-scale benchmark comparing random forest and logistic regression classifiers; methodological comparison of learning models.", "model": "claude-haiku-5-5"}
++{"work_id": 2895972434, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Meta-learning for recommending ensemble combination methods; automated ML design, fits ml methodology weakly.", "model": "claude-haiku-5-5"}
++{"work_id": 3126343538, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Framework for evaluating interpretability of deep models via consensus; interpretability evaluation as a safety/transparency problem.", "model": "claude-haiku-5-5"}
++{"work_id": 3000034219, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Book chapter on causal models for dynamical systems; causal inference is a core ml_theory topic.", "model": "claude-haiku-5-5"}
++{"work_id": 3012094357, "categories": ["ml_theory", "database_systems"], "primary": "ml_theory", "confidence": 0.4, "rationale": "One-class SVDD-based anomaly detection on operations and maintenance data; mainly a classical learning method applied to IT data, so ambiguous between ml_theory and systems/data management.", "model": "claude-haiku-5-5"}
++{"work_id": 4225697294, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.55, "rationale": "Classification of cloud images with deep neural networks for imbalanced data; image classification application with DL models.", "model": "claude-haiku-5-5"}
++{"work_id": 2897243396, "categories": ["scientific_computing", "algorithms"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Adaptive sparse grid least-squares regression with coordinate rotation preprocessing; numerical methods for regression.", "model": "claude-haiku-5-5"}
++{"work_id": 4409493966, "categories": ["cybersecurity", "ml_theory"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Malicious network anomaly detection with Prophet, Luminol and RNN to detect botnets and attacks.", "model": "claude-haiku-5-5"}
++{"work_id": 7166662462, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Theoretical analysis of the inlier-memorization effect in autoencoders for outlier detection.", "model": "claude-haiku-5-5"}
++{"work_id": 7133193233, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.75, "rationale": "Minimax mean estimation rates for high-dimensional Markov Gaussian mixture models; statistical learning theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4392903168, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Diffusion model applied to unsupervised anomaly detection in multivariate time series.", "model": "claude-haiku-5-5"}
++{"work_id": 4416873414, "categories": ["llm", "database_systems", "software_engineering"], "primary": "llm", "confidence": 0.35, "rationale": "Anomaly detection in Industry 4.0 with LLMs generating explanations; mixes classical detection with LLM explanation, reliability framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4379538576, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Hierarchical quadratic random forest classifier with penalized discriminants in decision nodes; ensemble classifier design.", "model": "claude-haiku-5-5"}
++{"work_id": 2928518543, "categories": ["nlp", "information_retrieval"], "primary": "nlp", "confidence": 0.4, "rationale": "Unsupervised analysis of natural language content on Reddit communities to detect anomalies and disturbance events; text-based community analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 7208751990, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.4, "rationale": "Real-time financial transaction monitoring with stream processing and distributed computing; ML for risk detection is applied, not a core category.", "model": "claude-haiku-5-5"}
++{"work_id": 4385346229, "categories": ["ml_theory", "reinforcement_learning"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Bayesian meta-learning with accelerated inference; closest to probabilistic/learning methodology, though it is not clearly a deep architecture or RL paper.", "model": "claude-haiku-5-5"}
++{"work_id": 2900419943, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Accelerated augmented Lagrangian method with iteration-complexity analysis for multi-criteria optimization; algorithm design and analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 7126019639, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Diffusion-based controllable medical anomaly image generation and detection; main contribution is the generative model, with detection/localization as downstream.", "model": "claude-haiku-5-5"}
++{"work_id": 4234642263, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Applies standard ML classifiers to credit card fraud detection; application-focused, weakly fits any category, most closely general ML methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 7132956912, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Dissertation deriving Rademacher generalization bounds for neural networks based on feature diversity.", "model": "claude-haiku-5-5"}
++{"work_id": 7213442406, "categories": ["ml_theory", "computational_biology"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Multi-objective hyperparameter selection and ensembling for imbalanced clinical prediction; methodology with a medical application.", "model": "claude-haiku-5-5"}
++{"work_id": 3027701685, "categories": ["cybersecurity", "database_systems"], "primary": null, "confidence": 0.25, "rationale": "Anomaly detection in application performance monitoring time series with ML; ambiguous fit, primarily systems monitoring rather than a clear ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 3173110490, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Variational inference for online Bayesian nonparametric clustering with control over over-clustering; probabilistic modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 4396611119, "categories": ["ml_theory"], "primary": null, "confidence": 0.2, "rationale": "Applies linear and other ML models to refine PV production forecasts; energy application with weak fit to any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4365509942, "categories": ["ml_theory", "formal_verification"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Supervised classification method using formal methods to generate formulas and models; primarily a learning technique, with formal methods as the core tool.", "model": "claude-haiku-5-5"}
++{"work_id": 3204747322, "categories": ["programming_languages", "formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Process algebra decomposition with bisimulation and state-space exploration, a formal specification and verification technique.", "model": "claude-haiku-5-5"}
++{"work_id": 3005800578, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Autobiographical account of logic and computational argumentation; primarily philosophy/informal logic rather than a CS research contribution, though argument mapping is adjacent.", "model": "claude-haiku-5-5"}
++{"work_id": 3165175988, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Model checking and simulation of DEVS activity models for verification and validation of behavioral specifications.", "model": "claude-haiku-5-5"}
++{"work_id": 2981823863, "categories": ["computer_architecture", "formal_verification"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Hardware SAT-solver design with clause learning in a standalone device, a digital hardware design contribution applied to verification.", "model": "claude-haiku-5-5"}
++{"work_id": 4403534925, "categories": [], "primary": null, "confidence": 0.35, "rationale": "Regression-based profit forecasting for dairy farm management; agricultural economics/operations rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 6927776925, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.7, "rationale": "Compositional conversion of LTLf formulas to DFAs for model checking and verification tooling.", "model": "claude-haiku-5-5"}
++{"work_id": 2904264076, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "First-principles computation of elastic properties of refractory alloys; materials science computation, not core CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4378551252, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.75, "rationale": "Automated proof of state-space equivalence for Petri nets using SMT encoding.", "model": "claude-haiku-5-5"}
++{"work_id": 4405491888, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.3, "rationale": "Colored Petri net modeling and simulation of drone delivery logistics; performance modeling with weak CS-theory fit, closest to simulation tooling.", "model": "claude-haiku-5-5"}
++{"work_id": 2985419616, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Philosophy/history of logic chapter on syllogistic and supposition theory; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4387635646, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Automata-based SMT string solver for word equations; automated reasoning and solving falls under formal verification.", "model": "claude-haiku-5-5"}
++{"work_id": 2998652541, "categories": ["computer_vision", "hci"], "primary": "hci", "confidence": 0.3, "rationale": "Smartphone indoor positioning via sensor fusion; closest fits are ubiquitous/mobile sensing (hci) or robotics-style localisation, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4416603322, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Automatic structures and algebraic characterisations of groups in logic-based complexity classes; closest to computability and automata theory, though it is mathematical logic.", "model": "claude-haiku-5-5"}
++{"work_id": 3041782238, "categories": ["deep_learning_architectures", "ai_safety", "formal_verification"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Adversarial robustness of quantized neural networks verified with SMT solvers; robustness is the central theme.", "model": "claude-haiku-5-5"}
++{"work_id": 7172406392, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.4, "rationale": "Symbolic execution and PRISM model checking for probabilistic programs; abstract is uninformative but title indicates verification.", "model": "claude-haiku-5-5"}
++{"work_id": 4292072122, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Max-plus polyhedra state estimation for uncertain max-plus linear systems; control/systems theory, closest to ml_theory filtering but weakly fitting.", "model": "claude-haiku-5-5"}
++{"work_id": 4210776678, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Satisfiability algorithms for string constraints and word equations via automata, automated reasoning.", "model": "claude-haiku-5-5"}
++{"work_id": 2919546362, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Improving indoor PDR trajectories from pedestrian sensor data for location-based services; ubiquitous computing framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4409052338, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.7, "rationale": "Improving CDCL SAT solver heuristics (restarts, branching); SAT solving falls under automated reasoning.", "model": "claude-haiku-5-5"}
++{"work_id": 4360995890, "categories": ["formal_verification", "software_engineering"], "primary": "formal_verification", "confidence": 0.7, "rationale": "Modeling CPS safety with AADL extension and TCSP formal verification via model transformation.", "model": "claude-haiku-5-5"}
++{"work_id": 4401567949, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Algorithm to identify independent variables in LTL specifications using model checkers.", "model": "claude-haiku-5-5"}
++{"work_id": 7202355676, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.3, "rationale": "Petri net model for fleet formation in logistics; operations research modeling, weakly fits any category.", "model": "claude-haiku-5-5"}
++{"work_id": 7167214329, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Bounded model checker for Rust with contracts and verification of safety properties.", "model": "claude-haiku-5-5"}
++{"work_id": 2941019358, "categories": ["formal_verification", "quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Circuit relations and ZX-calculus axiomatization for quantum stabilizer fragments.", "model": "claude-haiku-5-5"}
++{"work_id": 2888263960, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Materials science study of high-entropy nitride films; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3036333943, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.75, "rationale": "HOL Light theorem proving analysis of robotic cell injection dynamics.", "model": "claude-haiku-5-5"}
++{"work_id": 2980088699, "categories": ["formal_verification"], "primary": "programming_languages", "confidence": 0.4, "rationale": "Category-theoretic study of differential categories and tangent categories; mathematical foundations of programming semantics.", "model": "claude-haiku-5-5"}
++{"work_id": 2809414278, "categories": ["formal_verification", "database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Complex event processing over streams with temporal logic and timed automata pattern matching.", "model": "claude-haiku-5-5"}
++{"work_id": 2969344391, "categories": ["ml_theory", "software_engineering"], "primary": "software_engineering", "confidence": 0.35, "rationale": "Information model for computational accountability in distributed systems using ORM and ASP.", "model": "claude-haiku-5-5"}
++{"work_id": 2810295409, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Observability testing of DES specification languages via formal logic and automata; formal methods on discrete-event systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2942235760, "categories": ["formal_verification", "blockchain"], "primary": "formal_verification", "confidence": 0.85, "rationale": "Deductive verification of Ethereum smart contracts in Why3 is the main contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4414118222, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Constraint-solving algorithm modulo commutativity with fuzzy relations; soundness and termination proofs; algorithmic in nature.", "model": "claude-haiku-5-5"}
++{"work_id": 4416018542, "categories": ["llm", "formal_verification"], "primary": "llm", "confidence": 0.6, "rationale": "Evaluates LLM-based autoformalization and theorem proving pipelines on a benchmark; benchmark work on LLM reasoning in Lean.", "model": "claude-haiku-5-5"}
++{"work_id": 4391021938, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Control-theoretic interval observer design for hybrid systems; closest fit is theory of estimation, though it is systems and control rather than CS-core.", "model": "claude-haiku-5-5"}
++{"work_id": 4402325914, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Control barrier function safety controllers applied to autonomous trucks with experimental deployment on a full-scale vehicle.", "model": "claude-haiku-5-5"}
++{"work_id": 4416534401, "categories": ["blockchain", "formal_verification"], "primary": "formal_verification", "confidence": 0.85, "rationale": "Model checking with SPIN and LTL of Ethereum Beacon Chain epoch processing.", "model": "claude-haiku-5-5"}
++{"work_id": 3098615674, "categories": ["algorithms", "formal_verification"], "primary": "algorithms", "confidence": 0.5, "rationale": "Answer set programming optimization and lazy-grounding transformations; solver algorithm efficiency with verification of transformations.", "model": "claude-haiku-5-5"}
++{"work_id": 7128294598, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Supervisory control of airport traffic modeled with Petri nets to prevent collisions; discrete-event formal modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 4401630734, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.4, "rationale": "V-nets formalism for supervision and diagnosis of discrete event sequences in industrial processes; formalism-focused.", "model": "claude-haiku-5-5"}
++{"work_id": 3158571805, "categories": ["ai_agents", "robotics"], "primary": "robotics", "confidence": 0.3, "rationale": "Model predictive control for traffic management in smart cities with hybrid/event-driven systems; control-theoretic, weakly fits intelligent transport systems, not clearly a listed category.", "model": "claude-haiku-5-5"}
++{"work_id": 4415180303, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.45, "rationale": "Sim2Real gap quantified with neural network with formal guarantees; applied to Mecanum bot and pendulum controllers.", "model": "claude-haiku-5-5"}
++{"work_id": 4400077792, "categories": ["robotics"], "primary": "robotics", "confidence": 0.4, "rationale": "Midcourse-terminal interceptor guidance with reachability analysis; guidance of autonomous vehicles/missiles, closest to robotics autonomy.", "model": "claude-haiku-5-5"}
++{"work_id": 4396882310, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Semantics of resource logics (Bunched Implications, Linear Logic) for program reasoning; language semantics as design.", "model": "claude-haiku-5-5"}
++{"work_id": 7124134160, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Verification of opacity properties of discrete event systems via automata and diagnosers.", "model": "claude-haiku-5-5"}
++{"work_id": 2797906474, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.35, "rationale": "Petri net model extended for uncertain knowledge representation; modelling formalism for discrete event systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2947544026, "categories": ["formal_verification", "cybersecurity"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Runtime monitoring with metric temporal logic for intrusion detection policies; verification formalism applied to security.", "model": "claude-haiku-5-5"}
++{"work_id": 4391985801, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.4, "rationale": "New logic and assume-guarantee contracts to prove stability of layered control systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4391937166, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Formal language and verification of behavior trees used in robotics.", "model": "claude-haiku-5-5"}
++{"work_id": 4366551135, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.55, "rationale": "Logical framework extending LF with rational terms and coinduction; type theory and decidability.", "model": "claude-haiku-5-5"}
++{"work_id": 4406549421, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Lightweight network for low-light image enhancement; image restoration is computer vision, architecture is lightweight design.", "model": "claude-haiku-5-5"}
++{"work_id": 2999748008, "categories": ["nlp", "computer_vision"], "primary": "nlp", "confidence": 0.5, "rationale": "Translation of ancient script to Sinhala with character recognition; mixes OCR-style image processing with language translation.", "model": "claude-haiku-5-5"}
++{"work_id": 3206423720, "categories": ["ai_safety", "computer_vision"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Adversarial attack transferability on image classifiers, a robustness/security topic on DNNs.", "model": "claude-haiku-5-5"}
++{"work_id": 4403045668, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Aircraft detection in satellite imagery via region proposal and classification.", "model": "claude-haiku-5-5"}
++{"work_id": 7163998453, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Single-image dehazing network for visibility restoration.", "model": "claude-haiku-5-5"}
++{"work_id": 7213494877, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Image classification framework emphasizing training techniques and data augmentation.", "model": "claude-haiku-5-5"}
++{"work_id": 2996931328, "categories": ["deep_learning_architectures", "ai_accelerators"], "primary": "deep_learning_architectures", "confidence": 0.7, "rationale": "Efficient CNN architectures for inference with parameter and operation reduction (Micronet challenge).", "model": "claude-haiku-5-5"}
++{"work_id": 7164124236, "categories": ["multimodal_ai", "computer_graphics", "llm"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "VLMs generating Blender code for 3D reconstruction, combining vision-language models with 3D graphics.", "model": "claude-haiku-5-5"}
++{"work_id": 4404954354, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Medical ultrasound image classification of thyroid nodules comparing CNN, ViT and hybrid models.", "model": "claude-haiku-5-5"}
++{"work_id": 4298009864, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.8, "rationale": "E-commerce product recognition and segmentation with attention-augmented YOLACT.", "model": "claude-haiku-5-5"}
++{"work_id": 7203466192, "categories": ["foundation_models", "generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Controllable retinal image generation from foundation model latent spaces; generative synthesis evaluation on medical imaging.", "model": "claude-haiku-5-5"}
++{"work_id": 4417350489, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.75, "rationale": "Interpreting and modifying transformer self-attention architecture via image processing framework.", "model": "claude-haiku-5-5"}
++{"work_id": 4366827124, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Instance segmentation of uterine MRI images, a medical image analysis task.", "model": "claude-haiku-5-5"}
++{"work_id": 4400230393, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Video intra-coding resolution adaptation with CNN prediction; closest is image/video coding, arguably computer networks or multimedia, but CV-adjacent.", "model": "claude-haiku-5-5"}
++{"work_id": 7155178947, "categories": ["generative_ai", "computer_graphics"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Text-driven dance motion generation with a motion transformer and dataset.", "model": "claude-haiku-5-5"}
++{"work_id": 4402917109, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.75, "rationale": "Gaussian splatting for 3D reconstruction and novel view rendering.", "model": "claude-haiku-5-5"}
++{"work_id": 4381948429, "categories": ["computer_vision", "hci"], "primary": "computer_vision", "confidence": 0.6, "rationale": "YOLO-based person detection to automate classroom lights and fans.", "model": "claude-haiku-5-5"}
++{"work_id": 7112321703, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Augmented reality mobile shoe-fitting application and user requirements.", "model": "claude-haiku-5-5"}
++{"work_id": 2886945075, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.7, "rationale": "Text classification of newsgroups using CNN with GloVe embeddings.", "model": "claude-haiku-5-5"}
++{"work_id": 2950134147, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Gait-based age and gender estimation from video using CNNs.", "model": "claude-haiku-5-5"}
++{"work_id": 3215388118, "categories": ["computer_vision", "generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Uses a GAN for WiFi CSI activity recognition; device-free sensing with generative model, sensing-based human activity recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 4390361966, "categories": ["computer_vision", "cybersecurity"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Copy-move image forgery detection with CNN and feature selection; image forensics on images.", "model": "claude-haiku-5-5"}
++{"work_id": 4388667087, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Segmentation of choroidal vasculature in OCT volumetric medical images.", "model": "claude-haiku-5-5"}
++{"work_id": 3112061614, "categories": ["computer_vision", "information_retrieval"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Image classification with weighted nearest neighbor for tag recommendation on image sharing sites.", "model": "claude-haiku-5-5"}
++{"work_id": 4416442992, "categories": ["computer_vision", "generative_ai", "scientific_computing"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Diffusion-based generation of ground surfaces (DTM) from DSM; geospatial image-to-image translation.", "model": "claude-haiku-5-5"}
++{"work_id": 4392442054, "categories": ["cybersecurity", "computer_vision"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Digital watermarking of medical images for integrity and authenticity protection.", "model": "claude-haiku-5-5"}
++{"work_id": 2907431110, "categories": ["computer_networks", "computer_vision", "robotics"], "primary": "computer_networks", "confidence": 0.4, "rationale": "Visible light communication-based vehicle tracking using cameras; positioning and tracking system with communication signals.", "model": "claude-haiku-5-5"}
++{"work_id": 4386257148, "categories": ["computer_networks", "computational_biology", "edge_computing"], "primary": "edge_computing", "confidence": 0.4, "rationale": "IoT wearable architecture for Alzheimer's patient monitoring and caregiver platform; health informatics via IoT.", "model": "claude-haiku-5-5"}
++{"work_id": 2799322609, "categories": ["computer_vision", "hci"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Monocular camera-based 3D pose estimation with Mean-Shift tracking for virtual reality.", "model": "claude-haiku-5-5"}
++{"work_id": 3194681712, "categories": ["computer_vision", "hci"], "primary": "computer_vision", "confidence": 0.4, "rationale": "Radar-based gesture recognition and segmentation with CNNs; sensor-based recognition for interaction.", "model": "claude-haiku-5-5"}
++{"work_id": 4408017987, "categories": ["computer_vision", "generative_ai"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Digital image processing framework covering edge extraction, style transfer and compression; primarily image processing, with style transfer touching generative synthesis.", "model": "claude-haiku-5-5"}
++{"work_id": 7167467722, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.7, "rationale": "Tutorial-style overview of neural network architectures (RNN, LSTM, GRU, CNN) and training procedures.", "model": "claude-haiku-5-5"}
++{"work_id": 4416620171, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Deep learning on microscopy images of Gram stains for medical image analysis; medical image analysis is shared between computer vision and biomedical informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 4393212678, "categories": ["ai_safety", "computer_vision"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Analyses inconsistencies in saliency-map explanations for medical deep learning, a transparency/interpretability concern.", "model": "claude-haiku-5-5"}
++{"work_id": 2949391068, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Learned dynamic scaling policy inside CNNs evaluated on ImageNet, COCO and VOC for recognition and segmentation.", "model": "claude-haiku-5-5"}
++{"work_id": 4313830031, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Handcrafted multi-feature fusion (LBP, HOG, colour) for plant leaf image recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 7152933672, "categories": ["generative_ai", "multimodal_ai", "llm"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Survey of generative video trailer synthesis using LLMs, MLLMs and text-to-video models.", "model": "claude-haiku-5-5"}
++{"work_id": 4400934125, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Image classification for industrial images using multi-scale feature extraction with stochastic configuration networks.", "model": "claude-haiku-5-5"}
++{"work_id": 2968866686, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Attention-based fine-grained image recognition with CNNs on standard benchmarks.", "model": "claude-haiku-5-5"}
++{"work_id": 2968914950, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Image segmentation of psoriasis skin regions for medical diagnosis using colour-space analysis and k-means.", "model": "claude-haiku-5-5"}
++{"work_id": 3090375472, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Low-light image enhancement is image restoration/enhancement, a computer vision task using a learned network.", "model": "claude-haiku-5-5"}
++{"work_id": 4388642403, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Domain adaptation for 3D MRI classification; medical image analysis shared with computer vision, with biomedical application framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7199635077, "categories": ["hci", "computer_vision"], "primary": "hci", "confidence": 0.5, "rationale": "Elderly fall detection via passive Wi-Fi radar and ML is assistive technology; signal classification with AlexNet is secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 7140831868, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Real-time face detection and gender/age classification from CCTV images is visual recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 4415102501, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Flow-based generative image compression with tokenization; generative model is the core contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3116998496, "categories": ["robotics"], "primary": null, "confidence": 0.3, "rationale": "Gait acceleration sensor data analysis; weakly fits sensing and biomechanics, low confidence in any category.", "model": "claude-haiku-5-5"}
++{"work_id": 2769131622, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.8, "rationale": "CNN architecture for classifying diabetic foot ulcer images; image classification in a medical setting.", "model": "claude-haiku-5-5"}
++{"work_id": 4411552410, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Cattle face recognition with metric learning is image recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 4387664177, "categories": ["computer_vision", "computer_graphics"], "primary": "computer_vision", "confidence": 0.55, "rationale": "AR camera pose registration is vision-based tracking; AR links to computer graphics and HCI.", "model": "claude-haiku-5-5"}
++{"work_id": 3010412574, "categories": ["ml_theory", "computer_vision"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Parameter tuning in meta-learning and zero-shot learning for image classification; methodological focus on learning.", "model": "claude-haiku-5-5"}
++{"work_id": 4410134551, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Non-technical book on AI's role in society and business; not a CS research contribution in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 7168123153, "categories": ["nlp"], "primary": null, "confidence": 0.3, "rationale": "Systematic review of AI use in mathematics education; education-domain study with weak fit to any CS category, so conservatively tagged only as nlp-adjacent is doubtful; better left empty.", "model": "claude-haiku-5-5"}
++{"work_id": 3168036902, "categories": ["computer_graphics"], "primary": null, "confidence": 0.3, "rationale": "Turbomachinery blade inverse design with PSO; engineering design optimisation, closest to scientific_computing but not clearly CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4318620721, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Automated symbolic-regression discovery of catalytic kinetic models for chemical reaction engineering; scientific modelling tooling.", "model": "claude-haiku-5-5"}
++{"work_id": 6929444221, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Gaussian constitutive neural networks for material modelling with physics-informed discovery; scientific machine learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7169848785, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Preface with no abstract; cannot determine a category from available metadata.", "model": "claude-haiku-5-5"}
++{"work_id": 7156328285, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Genetic programming symbolic regression for system identification of dynamical systems under noise; scientific modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 4414898387, "categories": ["autonomous_scientific_discovery", "generative_ai"], "primary": "autonomous_scientific_discovery", "confidence": 0.7, "rationale": "Closed-loop generative and active learning workflow driving molecular materials discovery for carbon capture.", "model": "claude-haiku-5-5"}
++{"work_id": 3003447096, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Data conversion and repository architecture for materials characterization data; data management focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4410289386, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Neural network representation of effective interactions for coupled-cluster many-body chemistry models.", "model": "claude-haiku-5-5"}
++{"work_id": 7162445105, "categories": ["ai_agents", "autonomous_scientific_discovery", "computational_biology"], "primary": "autonomous_scientific_discovery", "confidence": 0.8, "rationale": "Agent harness for autonomous atomistic research with AI scientists; spans materials, chemistry and drug discovery simulation workflows.", "model": "claude-haiku-5-5"}
++{"work_id": 4405426097, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Formal behavioural distance measure between mathematical expressions via Wasserstein distance; closest to theory/methodology, though it touches symbolic regression and scientific_computing.", "model": "claude-haiku-5-5"}
++{"work_id": 4402242289, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Review of ML and DL models for predicting chemical carcinogenicity, a toxicology/health informatics application.", "model": "claude-haiku-5-5"}
++{"work_id": 7113050848, "categories": ["ai_agents", "autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.55, "rationale": "Demonstrates Deep Research AI tools conducting autonomous research and studies.", "model": "claude-haiku-5-5"}
++{"work_id": 7164296032, "categories": ["deep_learning_architectures", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Topology-aware message-passing GNN with persistent homology; applied to MOF property prediction.", "model": "claude-haiku-5-5"}
++{"work_id": 7214605398, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Psychometric instrument for teachers' beliefs about AI; human-AI attitudes, education-focused, not a clear CS fit beyond human-AI interaction.", "model": "claude-haiku-5-5"}
++{"work_id": 4399747476, "categories": [], "primary": null, "confidence": 0.3, "rationale": "General survey-style overview of AI trends, startups and Industry 4.0 with no single clear technical contribution category.", "model": "claude-haiku-5-5"}
++{"work_id": 7171536644, "categories": ["ai_agents", "autonomous_scientific_discovery", "reinforcement_learning"], "primary": "autonomous_scientific_discovery", "confidence": 0.7, "rationale": "Evaluates search efficiency of autonomous research systems and proposes fluid search with a portfolio bandit.", "model": "claude-haiku-5-5"}
++{"work_id": 4412068170, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.6, "rationale": "Classification of AI roles in chemistry up to AI Researcher generating hypotheses and autonomous design.", "model": "claude-haiku-5-5"}
++{"work_id": 7113026192, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "AI-powered personalized learning platform evaluated in elementary math education; user-facing educational technology study.", "model": "claude-haiku-5-5"}
++{"work_id": 7117871859, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Position paper on AI tools in mathematics education; education research rather than a CS category, so no category assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 7139945210, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Pure number theory / algebraic geometry on R-equivalence of cubic surfaces; not a CS category, though it uses generative AI as a research tool.", "model": "claude-haiku-5-5"}
++{"work_id": 4411799567, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Editorial on AI in the research process and publishing; scientometrics/meta-science rather than a CS research category, with slight relation to autonomous scientific discovery.", "model": "claude-haiku-5-5"}
++{"work_id": 4414528656, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Qualitative language-education study on YouTube listening practice; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7154965180, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Mechanics of mushroom-based foods with constitutive model discovery; materials/soft-matter science, not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4403160419, "categories": ["deep_learning_architectures", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Equivariant SO(3) feature design for ML models; applied to molecular quantum properties, so chemistry application is secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4378470143, "categories": ["ml_theory", "scientific_computing"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Neural-network discovery of continuous symmetry groups and Lie algebras; learning-theoretic/mathematical methodology applied to physics.", "model": "claude-haiku-5-5"}
++{"work_id": 2896073529, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Announcement of a student essay contest; not a research paper.", "model": "claude-haiku-5-5"}
++{"work_id": 4226190677, "categories": ["scientific_computing", "computer_architecture"], "primary": "scientific_computing", "confidence": 0.45, "rationale": "Inverse design of metasurfaces with optimization and ML surrogates; electromagnetic engineering with ML tooling, closest to scientific computing, weakly to hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 7207633319, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Descriptive qualitative study of AI use in primary science education; education research, not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7127595070, "categories": ["nlp"], "primary": null, "confidence": 0.2, "rationale": "Scientific writing education workshop; not clearly a CS category. Uses AI-generated questions only incidentally; low-confidence fit to none.", "model": "claude-haiku-5-5"}
++{"work_id": 7164884815, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.35, "rationale": "AI Index report covering evaluation, governance, safety, and economics of AI broadly; general AI overview with governance emphasis, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7155543779, "categories": ["computer_vision", "deep_learning_architectures", "scientific_computing"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Vision Transformer architecture for XRD pattern image classification; applied to materials characterization.", "model": "claude-haiku-5-5"}
++{"work_id": 4385567506, "categories": ["scientific_computing", "reinforcement_learning"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Benchmark toolkit for inverse design of nanophotonic devices with open-source simulator; design algorithms compared.", "model": "claude-haiku-5-5"}
++{"work_id": 3131458733, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Information systems study of online learning adoption using SEM and neural networks; survey-based education research, not a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4406949642, "categories": ["autonomous_scientific_discovery", "scientific_computing"], "primary": "autonomous_scientific_discovery", "confidence": 0.4, "rationale": "AI-accelerated high-throughput materials discovery and synthesis for electronic materials; magazine overview.", "model": "claude-haiku-5-5"}
++{"work_id": 7121660003, "categories": ["information_retrieval", "ml_theory"], "primary": null, "confidence": 0.25, "rationale": "AutoML benchmarking on business analytics datasets with H2O; applied ML rather than a clear category.", "model": "claude-haiku-5-5"}
++{"work_id": 4391244979, "categories": ["generative_ai", "scientific_computing"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Self-conditioned diffusion model generating voxel mechanical metamaterials for inverse design.", "model": "claude-haiku-5-5"}
++{"work_id": 7204883563, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.75, "rationale": "LLM agent harness toolkit for data-science automation with benchmarks and agent operator programs.", "model": "claude-haiku-5-5"}
++{"work_id": 7163871521, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Science policy analysis of open science and data ownership; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4416200173, "categories": ["computer_graphics", "generative_ai", "deep_learning_architectures", "scientific_computing"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "Topology optimization of metamaterial microstructures on grid domains with CNN surrogate for stiffness; geometric design and fabrication fits graphics/computational design, with ML surrogate and scientific computing aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 4389566447, "categories": ["reinforcement_learning", "scientific_computing", "ml_theory"], "primary": "scientific_computing", "confidence": 0.55, "rationale": "Symbolic regression for physics using deep RL with units constraints to discover physical laws; automated discovery of equations fits scientific discovery and RL methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4200490276, "categories": ["ml_theory", "database_systems"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Automated machine learning pipeline search via beam search applied to power industry; AutoML is a methodology, with application focus. Conservative assignment.", "model": "claude-haiku-5-5"}
++{"work_id": 7165673775, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Pedagogical study of engineering education with digital technologies; not computer-science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7211978759, "categories": ["llm", "ai_safety"], "primary": "llm", "confidence": 0.7, "rationale": "Audits frontier LLMs for retrieval of memorized molecular property values; evaluation of LLMs with concerns about contamination and reliability.", "model": "claude-haiku-5-5"}
++{"work_id": 7170193477, "categories": ["database_systems", "cybersecurity"], "primary": "database_systems", "confidence": 0.4, "rationale": "LIMS architecture with workflow state model and data governance; systems architecture, with security/privacy aspects in HIPAA context.", "model": "claude-haiku-5-5"}
++{"work_id": 3214987860, "categories": ["computational_biology", "ml_theory"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Data augmentation library for chemical structures for property prediction with GNNs; chemistry/materials informatics with ML augmentation methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4416999558, "categories": ["ml_theory", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Explainable few-shot AI4E framework combining symbolic regression and physics for materials engineering constitutive modeling.", "model": "claude-haiku-5-5"}
++{"work_id": 7217524975, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Development of computer-assisted instruction for PowerPoint education study; educational research, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4404678493, "categories": ["multimodal_ai", "computational_biology", "scientific_computing"], "primary": "multimodal_ai", "confidence": 0.5, "rationale": "Multimodal neural network integrating image and text data to predict nanoparticle properties; materials simulation application.", "model": "claude-haiku-5-5"}
++{"work_id": 4224247694, "categories": ["deep_learning_architectures", "robotics"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Physical reservoir computing in adaptive metastructures; reservoir computing is a neural/learning framework applied to mechanical wave control, with adaptive structures leaning toward robotics/materials, so classification is uncertain.", "model": "claude-haiku-5-5"}
++{"work_id": 4410956047, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.3, "rationale": "Photovoltaic power forecasting with EEMD and fuzzy cognitive maps; an applied energy forecasting method with weak fit to any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4415102374, "categories": ["deep_learning_architectures", "sustainable_computing"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Algebraic weight computation for single-layer neural networks as an alternative to gradient descent, with energy-efficiency framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3011306938, "categories": ["information_retrieval", "database_systems"], "primary": "database_systems", "confidence": 0.35, "rationale": "Link prediction on n-ary relational master data with GNNs, anonymization and verification; data-management framing.", "model": "claude-haiku-5-5"}
++{"work_id": 2978220519, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Reservoir computing (echo state network) studied for reconstructing generalized synchronization in chaotic systems; nonlinear dynamics with learning model.", "model": "claude-haiku-5-5"}
++{"work_id": 7197424828, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.7, "rationale": "Training methods for quantized and binarized neural networks, including stochastic quantization and probabilistic propagation, are neural network training and compression.", "model": "claude-haiku-5-5"}
++{"work_id": 4312870744, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Fast multiplier hardware design for neural network implementation is hardware for efficient ML computation.", "model": "claude-haiku-5-5"}
++{"work_id": 4403331393, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.35, "rationale": "Optimal power flow solved with PSO combined with ANN; power system optimization with weak fit to ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 7118589965, "categories": ["deep_learning_architectures", "computer_vision", "edge_computing"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Channel pruning and few-shot meta-learning for compact models on edge devices for plant disease images; primarily model compression applied to vision.", "model": "claude-haiku-5-5"}
++{"work_id": 4407405188, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.45, "rationale": "LSTM-based interacting multiple model target tracking; deep learning architecture for state estimation, with tracking as the application.", "model": "claude-haiku-5-5"}
++{"work_id": 7126095475, "categories": ["ml_theory", "database_systems"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Comparative ML/DL dropout prediction on institutional data; applied education analytics, weakly fits ML methodology; no clear ontology match beyond general ML evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 4391892579, "categories": ["information_retrieval", "database_systems"], "primary": "information_retrieval", "confidence": 0.45, "rationale": "Hypergraph learning over POI data with query-guided retrieval of strategic groups; closest to search/recommendation and data-mining over POI data.", "model": "claude-haiku-5-5"}
++{"work_id": 4413152908, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Bias-variance-correlation tradeoff for ML in science; statistical learning theory, though the abstract is a slide-deck fragment.", "model": "claude-haiku-5-5"}
++{"work_id": 4400948005, "categories": ["nlp", "reinforcement_learning"], "primary": "nlp", "confidence": 0.2, "rationale": "Deep learning for crypto quantitative trading using sentiment; finance application with weak CS category fit, sentiment from text is the only NLP link.", "model": "claude-haiku-5-5"}
++{"work_id": 3042576877, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.55, "rationale": "Learning rules and objective functions for PCA convergence; optimization theory for learning.", "model": "claude-haiku-5-5"}
++{"work_id": 4416583002, "categories": ["computer_vision", "scientific_computing"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Super-resolution of biomass maps from satellite imagery using a transformer; remote-sensing image enhancement.", "model": "claude-haiku-5-5"}
++{"work_id": 4408447422, "categories": ["nlp", "software_engineering"], "primary": "nlp", "confidence": 0.35, "rationale": "Deep learning for automatic English essay scoring and writing improvement; NLP evaluation framed in education.", "model": "claude-haiku-5-5"}
++{"work_id": 4296588701, "categories": ["edge_computing", "deep_learning_architectures"], "primary": "edge_computing", "confidence": 0.3, "rationale": "Emergency mobility prediction under IoT availability; graph neural networks on traffic data, IoT data quality and resources.", "model": "claude-haiku-5-5"}
++{"work_id": 4402273825, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Stability criterion for delayed neural networks via Lyapunov-Krasovskii functionals; theoretical analysis of neural network dynamics.", "model": "claude-haiku-5-5"}
++{"work_id": 2797984158, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Semi-supervised multi-kernel extreme learning machine; kernel methods and learning algorithm for SLFNs.", "model": "claude-haiku-5-5"}
++{"work_id": 7141530726, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.3, "rationale": "Battery SoC estimation combining electrochemical model with a neural network; not clearly in ontology. Closest is a hybrid neural network approach, weakly applicable; low confidence. Battery electrochemistry is outside listed categories.", "model": "claude-haiku-5-5"}
++{"work_id": 7148783783, "categories": ["computer_vision", "deep_learning_architectures", "quantum_computing"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Thesis on handwriting recognition with CNNs, spatial transformer networks, recurrent architectures and a hybrid classical-quantum network; main contributions are architectural with handwriting recognition as application.", "model": "claude-haiku-5-5"}
++{"work_id": 4416177815, "categories": ["deep_learning_architectures", "scientific_computing"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Hybrid VMD-KAN-Informer forecasting model; primarily a deep architecture combination applied to railway load forecasting, which is an energy/engineering application.", "model": "claude-haiku-5-5"}
++{"work_id": 3159894882, "categories": ["distributed_computing", "deep_learning_architectures"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "GPU-accelerated graph sampling for training graph representation learning; parallel systems for ML training.", "model": "claude-haiku-5-5"}
++{"work_id": 4391810135, "categories": ["hci", "computational_biology"], "primary": "computational_biology", "confidence": 0.35, "rationale": "Model of finger coordination in piano keystrokes using motion capture and BP neural network; human motor modelling, closest to biological/human-movement modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 3031233905, "categories": ["computer_architecture", "speech_audio", "ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.45, "rationale": "Spin-wave reservoir computing hardware evaluated on spoken digit recognition; hardware-based neuromorphic computing for ML.", "model": "claude-haiku-5-5"}
++{"work_id": 4391129611, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Physics-informed neural network applied to compressible gas pipeline transient simulation, compared with finite difference methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4407360793, "categories": ["scientific_computing", "deep_learning_architectures"], "primary": "scientific_computing", "confidence": 0.55, "rationale": "Physics-informed graph neural network as fast alternative to power flow solvers; scientific ML for power systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4393372557, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.45, "rationale": "Hyperdimensional classification of brain graphs from fMRI data; neuroimaging analysis application.", "model": "claude-haiku-5-5"}
++{"work_id": 4387116566, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Recurrent survival deep learning for cancer incidence prediction from biobank data; biomedical and health informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 4391129094, "categories": ["deep_learning_architectures", "scientific_computing"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Genetic-algorithm-optimized BP neural network applied to chemical reaction data; mainly a neural network methodology with a scientific application.", "model": "claude-haiku-5-5"}
++{"work_id": 4414859951, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.4, "rationale": "Reservoir computing demonstrated in a physical photonic resonator; hardware-based unconventional computing for ML, though abstract is nearly empty.", "model": "claude-haiku-5-5"}
++{"work_id": 7161991053, "categories": ["deep_learning_architectures", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Compares LSTM and Transformer sequence models on simulated healthcare claims event sequences; applied classification with a biomedical focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3195270160, "categories": ["algorithms", "deep_learning_architectures"], "primary": "algorithms", "confidence": 0.6, "rationale": "Applies GNNs and other learning models as heuristics for the NP-hard Steiner tree problem, a combinatorial optimization task.", "model": "claude-haiku-5-5"}
++{"work_id": 4386990763, "categories": ["information_retrieval", "deep_learning_architectures"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Knowledge-graph-based personalized attraction recommender, a recommender system contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7220870911, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Overview article on artificial intelligence history and education in Uzbek; not a specific CS research contribution, mostly general survey of AI topics.", "model": "claude-haiku-5-5"}
++{"work_id": 7109160285, "categories": ["reinforcement_learning", "ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Variational principle deriving softmax gating and energy-based dynamics for behavioral composition; theoretical modelling with agent applications.", "model": "claude-haiku-5-5"}
++{"work_id": 4378192806, "categories": ["deep_learning_architectures", "information_retrieval"], "primary": "deep_learning_architectures", "confidence": 0.3, "rationale": "Clustering-based LSTM forecasting of dengue time series; applied time-series forecasting with epidemiological focus, weakly fitting any category.", "model": "claude-haiku-5-5"}
++{"work_id": 4410775015, "categories": ["computer_vision", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "CNN breast cancer prediction on tabular Wisconsin dataset; medical diagnosis application of neural networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4407130511, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.7, "rationale": "Thesis on graph neural differential equation networks improving GNN representation learning and robustness.", "model": "claude-haiku-5-5"}
++{"work_id": 3047152076, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Theoretical physics-style analysis of neural network learning dynamics as a model of the universe; mainly theoretical treatment of trainable and hidden variables, borderline for ML theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4380551192, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Studies parameter-space symmetries and functional dimension of ReLU network architectures, a theoretical property of network architectures.", "model": "claude-haiku-5-5"}
++{"work_id": 4405619363, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Cultural assessment of generative AI systems for music creation; not a technical CS contribution but concerns generative AI and its cultural impact.", "model": "claude-haiku-5-5"}
++{"work_id": 7149698496, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Graph-based multi-channel fault diagnosis from sensor signals; a graph neural architecture applied to industrial fault classification, closest to architecture design with an application.", "model": "claude-haiku-5-5"}
++{"work_id": 2802497406, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.9, "rationale": "Neural network compression through sparseness allocation and neuron pruning.", "model": "claude-haiku-5-5"}
++{"work_id": 4408222477, "categories": ["federated_learning", "reinforcement_learning"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Adaptive learning rate scheduling for federated learning using PID control; primarily federated optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 4283646524, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Deep learning classification of endoscopic images for esophageal cancer diagnosis; medical image analysis, with clinical validation.", "model": "claude-haiku-5-5"}
++{"work_id": 4389157323, "categories": ["deep_learning_architectures", "scientific_computing"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Wind speed time-series forecasting combining denoising with TCN and GRU; a sequence model applied to energy forecasting.", "model": "claude-haiku-5-5"}
++{"work_id": 2907877223, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Multi-reservoir echo state network architecture for fuel cell remaining useful life prediction.", "model": "claude-haiku-5-5"}
++{"work_id": 4414684252, "categories": ["deep_learning_architectures", "scientific_computing"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Hypernetwork architecture for parametric dynamical system forecasting, with a scientific ML flavor.", "model": "claude-haiku-5-5"}
++{"work_id": 3183119475, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Co-evolution of opinion and signed-network dynamics with theoretical convergence results; closest to theory of dynamical models, though it is social-network dynamics rather than ML.", "model": "claude-haiku-5-5"}
++{"work_id": 2898975570, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.85, "rationale": "Analysis of multiway Quickselect variants with comparison and memory-transfer counts; algorithm design and analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 2895538521, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Spectral graph theory defining conjugate Laplacian matrices; graph algorithms/combinatorics, closest category is algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 4401726474, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Theoretical recovery conditions for block-sparse tensor recovery via OMP; signal-processing theory, closest to learning/statistical theory, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4400170167, "categories": ["database_systems", "information_retrieval"], "primary": "information_retrieval", "confidence": 0.45, "rationale": "Search algorithm for retrieval efficiency in large-scale datasets; vague abstract, retrieval framing with database scale.", "model": "claude-haiku-5-5"}
++{"work_id": 4385209635, "categories": ["quantum_computing", "algorithms"], "primary": "quantum_computing", "confidence": 0.55, "rationale": "QUBO formulations for Ising machines, combinatorial optimisation hardware; Ising machines are quantum/physics-inspired computing.", "model": "claude-haiku-5-5"}
++{"work_id": 7115725860, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.65, "rationale": "Online algorithms, competitive analysis and routing/scheduling optimisation for dynamic problems.", "model": "claude-haiku-5-5"}
++{"work_id": 2998687774, "categories": ["algorithms", "database_systems"], "primary": "algorithms", "confidence": 0.6, "rationale": "Textbook on theoretical algorithms for Big Data (streaming, sketching), so algorithms is primary.", "model": "claude-haiku-5-5"}
++{"work_id": 4205264571, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.8, "rationale": "Single-sample prophet inequalities and online/secretary-style algorithm design from SODA proceedings.", "model": "claude-haiku-5-5"}
++{"work_id": 4313546928, "categories": ["algorithms", "deep_learning_architectures"], "primary": "algorithms", "confidence": 0.5, "rationale": "Evolutionary dynamic multiobjective optimisation with neural-network prediction; metaheuristic focus gives algorithms as primary.", "model": "claude-haiku-5-5"}
++{"work_id": 3040673611, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Memetic swarm optimizer for large-scale water network optimization; metaheuristic algorithm design, application domain is engineering optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 3162282920, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.55, "rationale": "Applies Harris Hawks metaheuristic to Hyper Sudoku; combinatorial optimization/metaheuristics.", "model": "claude-haiku-5-5"}
++{"work_id": 4415799037, "categories": ["algorithms", "scientific_computing"], "primary": "algorithms", "confidence": 0.6, "rationale": "Local solvers for graph diffusion equations with sublinear-time algorithms; primarily algorithmic, with numerical solver framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4415332021, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.55, "rationale": "Sparse graph construction and seriation with provable query bounds; algorithm design for ordering problems.", "model": "claude-haiku-5-5"}
++{"work_id": 3197190047, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.5, "rationale": "Robustness and transmission capacity of heterogeneous networks optimized via MOEA; networking framed problem.", "model": "claude-haiku-5-5"}
++{"work_id": 7163898549, "categories": ["computational_complexity", "algorithms"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Addition-multiplication chains bounds in number theory; relates to shortest chain complexity, pure math combinatorics.", "model": "claude-haiku-5-5"}
++{"work_id": 2890999007, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Quantum-inspired local search enhanced binary PSO; metaheuristic algorithm design.", "model": "claude-haiku-5-5"}
++{"work_id": 7218127097, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Three-distance domination number of corona graphs; graph theory combinatorics, closest to algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 4400392879, "categories": [], "primary": null, "confidence": 0.4, "rationale": "General popular-science text on the history of algorithms; insufficient information for a CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 3151742782, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Dissertation on approximation algorithms for stochastic ranking, routing and classification.", "model": "claude-haiku-5-5"}
++{"work_id": 2911086832, "categories": ["algorithms", "computer_architecture"], "primary": null, "confidence": 0.2, "rationale": "Heuristic optimizer applied to PID tuning of an AVR control system; metaheuristic design belongs to algorithms, control application is outside the ontology, so low-confidence fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3200096258, "categories": ["computational_complexity"], "primary": null, "confidence": 0.2, "rationale": "Algebraic geometry of tensor varieties and degree generating functions; mathematics rather than CS, closest is algorithmic/algebraic computation with weak fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3136531308, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.7, "rationale": "Computes continuous mean distance of weighted graphs with discrete algorithms and computational geometry methods.", "model": "claude-haiku-5-5"}
++{"work_id": 7130513340, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "NP-completeness result and bounds on rainbow total domination in graphs, combinatorial hardness focus.", "model": "claude-haiku-5-5"}
++{"work_id": 6947983997, "categories": ["reinforcement_learning", "algorithms"], "primary": "reinforcement_learning", "confidence": 0.5, "rationale": "Online learning pacing strategy for repeated karma auctions with asymptotic optimality and Nash equilibrium analysis; learning-to-bid framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4392544619, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Matrix reversal operation description; mathematics rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3180083090, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Novel evolutionary estimation-of-distribution algorithm applied to smart grid competition benchmarks.", "model": "claude-haiku-5-5"}
++{"work_id": 4254111132, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.45, "rationale": "Tutorial on representations for evolutionary algorithms, a metaheuristic design topic.", "model": "claude-haiku-5-5"}
++{"work_id": 2981962366, "categories": ["hci"], "primary": null, "confidence": 0.3, "rationale": "Empirical study of students' learning strategies in computing education; education rather than a listed category, weak fit to hci.", "model": "claude-haiku-5-5"}
++{"work_id": 4414079667, "categories": ["algorithms", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Kriging surrogate models and improved simulated annealing for structural lightweight design of a tunnel shield via finite element analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4396597290, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Adaptive dynamic programming for Nash equilibrium guidance in a cooperative differential game; RL-style control applied to hypersonic vehicle guidance.", "model": "claude-haiku-5-5"}
++{"work_id": 4409058110, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Review of remaining useful life prediction for rotating machinery; industrial prognostics rather than a core CS category, so no category assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 3044771921, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Computing education study of teachers' views on contexts for teaching algorithms; it is education research, not a listed CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 3087997050, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Comparative study of metaheuristic optimization algorithms on benchmark functions.", "model": "claude-haiku-5-5"}
++{"work_id": 7203700339, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.95, "rationale": "W[1]-hardness and ETH lower bounds for norm maximization over zonotopes, a parameterized complexity result.", "model": "claude-haiku-5-5"}
++{"work_id": 4410943262, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Multi-objective metaheuristic optimization of a chemical reactor model in process simulation software.", "model": "claude-haiku-5-5"}
++{"work_id": 2758153130, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Monte Carlo confidence intervals for stochastic dynamic programming applied to option pricing; numerical methods for financial models.", "model": "claude-haiku-5-5"}
++{"work_id": 4404370937, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Differential evolution with adaptive gradient-based mutation for constrained optimization, applied to IoT device layout.", "model": "claude-haiku-5-5"}
++{"work_id": 7161020069, "categories": ["algorithms", "computational_complexity"], "primary": "computational_complexity", "confidence": 0.7, "rationale": "Welfare-proportionality tradeoff in temporal voting with NP-completeness, APX-hardness and fixed-parameter algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 2978556928, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.7, "rationale": "Metaheuristic Chemical Reaction Optimization applied to finding maximum independent set in graphs.", "model": "claude-haiku-5-5"}
++{"work_id": 2974268229, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.8, "rationale": "Local routing and weight-bounded geometric graph construction is algorithm design and analysis for geometric graphs.", "model": "claude-haiku-5-5"}
++{"work_id": 3035652685, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.45, "rationale": "Graph labeling (signed product cordial) is combinatorial mathematics, closest to algorithms/combinatorics; weak fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3123947211, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Smoothing functions for SOC complementarity problems are optimization methods; closest category is algorithms, with scientific_computing as an alternative.", "model": "claude-haiku-5-5"}
++{"work_id": 4320041016, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Double Roman domination number of middle graphs is graph-theoretic combinatorics tied to algorithmic bounds.", "model": "claude-haiku-5-5"}
++{"work_id": 4415250135, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.35, "rationale": "Antenna array placement for localization is signal/wireless sensing; closest ontology fit is hardware-related, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7024143063, "categories": ["algorithms", "computational_complexity"], "primary": "algorithms", "confidence": 0.7, "rationale": "Quasi-polynomial algorithms for Independent Set in hereditary graph classes, plus structural graph results on minimal separators.", "model": "claude-haiku-5-5"}
++{"work_id": 2982343211, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.55, "rationale": "Exact algorithm (MILP reduction) for arrow placement in graph drawings; graph drawing optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 3115336282, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Interior-point algorithm for linear optimization with complexity analysis; optimization algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 3166282595, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.35, "rationale": "Source code for estimating metaheuristic ability; metaheuristics fall under algorithms, low detail available.", "model": "claude-haiku-5-5"}
++{"work_id": 4413144369, "categories": ["algorithms", "ai_accelerators"], "primary": "algorithms", "confidence": 0.5, "rationale": "Compressed QUBO format for TSP formulations targeting Ising machines; combinatorial optimization with hardware motivation.", "model": "claude-haiku-5-5"}
++{"work_id": 7115594566, "categories": ["llm", "nlp", "database_systems"], "primary": "llm", "confidence": 0.6, "rationale": "Uses LLMs via prompt engineering to automatically build knowledge graphs/ontologies from news; LLM-centred with ontology/knowledge-graph construction (nlp/information extraction).", "model": "claude-haiku-5-5"}
++{"work_id": 4399491257, "categories": ["llm", "hci"], "primary": "llm", "confidence": 0.6, "rationale": "Custom-trained LLM chatbot as open educational resource, evaluated via student survey; LLM application plus user evaluation in education.", "model": "claude-haiku-5-5"}
++{"work_id": 7113152375, "categories": ["llm", "information_retrieval"], "primary": "llm", "confidence": 0.6, "rationale": "RAG-based academic chatbot using an LLM with document semantic search; system design centred on LLM retrieval augmentation.", "model": "claude-haiku-5-5"}
++{"work_id": 4402670518, "categories": ["nlp", "generative_ai"], "primary": "nlp", "confidence": 0.5, "rationale": "Generates synthetic annotated texts from knowledge graphs with LLMs to improve information extraction; NLP task focus with synthetic text generation.", "model": "claude-haiku-5-5"}
++{"work_id": 7149228871, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Supplementary data for a paper on distinguishing human vs LLM metaphors; minimal abstract, borderline NLP resource, conservatively rated as nlp-related but insufficient information.", "model": "claude-haiku-5-5"}
++{"work_id": 4416889670, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Non-scientific opinion/promotional text about a 'Future Knowledge Era' with no clear computer-science research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7221549073, "categories": ["ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Mechanistic analysis of jailbreak and over-refusal pathways in safety-aligned LLMs, introducing a safety routing method.", "model": "claude-haiku-5-5"}
++{"work_id": 7161067881, "categories": ["llm", "information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Benchmark and structured belief store for retrieval precision in LLM memory systems; retrieval-focused evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 7131836878, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.7, "rationale": "Prompt-engineering framework for LLM math word problem solving; reasoning with LLMs on an NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 4387799763, "categories": ["llm", "ai_safety", "nlp"], "primary": "llm", "confidence": 0.7, "rationale": "Evaluates ChatGPT self-explanations for sentiment analysis faithfulness and interpretability in LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4409475173, "categories": ["llm", "information_retrieval", "nlp"], "primary": "llm", "confidence": 0.8, "rationale": "RAG-based LLM educational chatbot; LLM with retrieval augmentation, also IR component.", "model": "claude-haiku-5-5"}
++{"work_id": 4394729406, "categories": ["nlp", "information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "BERT-based keyword search and movie retrieval using reviews; mixes NLP and search.", "model": "claude-haiku-5-5"}
++{"work_id": 4416454327, "categories": ["llm", "ai_safety"], "primary": "llm", "confidence": 0.7, "rationale": "Benchmark and evaluation of cultural competence in LLMs; alignment-related evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 7138061291, "categories": ["llm", "software_engineering"], "primary": "llm", "confidence": 0.75, "rationale": "Prompting strategies for LLM code generation with self-repair and unit tests.", "model": "claude-haiku-5-5"}
++{"work_id": 3109984837, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Palmprint biometric identification from images; image-based recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 4404852650, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Chinese paper text classification with BERT and hybrid deep learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7165886470, "categories": ["llm", "deep_learning_architectures"], "primary": "llm", "confidence": 0.75, "rationale": "Looped language model architecture and supervision analysis of transformers.", "model": "claude-haiku-5-5"}
++{"work_id": 7128818302, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.75, "rationale": "Fine-tuned T5 abstractive summarization of social media text; compares to ChatGPT.", "model": "claude-haiku-5-5"}
++{"work_id": 7214900323, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Survey study of language students' LLM use in foreign language education; primarily education research rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4414501685, "categories": ["nlp", "llm", "hci"], "primary": "nlp", "confidence": 0.6, "rationale": "LLM-based pseudo-labeling of product reviews for text classification; design-oriented text mining.", "model": "claude-haiku-5-5"}
++{"work_id": 4416245062, "categories": ["llm", "deep_learning_architectures"], "primary": "llm", "confidence": 0.8, "rationale": "Speculative decoding for LLM inference with online draft model selection; efficient inference of LLMs with bandit-based learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7117738325, "categories": ["llm", "autonomous_scientific_discovery", "reinforcement_learning"], "primary": "autonomous_scientific_discovery", "confidence": 0.75, "rationale": "Training LLMs as AI co-scientists to generate research plans via RL with self-grading rubrics.", "model": "claude-haiku-5-5"}
++{"work_id": 4391976463, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.8, "rationale": "Human evaluation of ChatGPT machine translation quality across domains.", "model": "claude-haiku-5-5"}
++{"work_id": 7110969657, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.6, "rationale": "Evaluates ChatGPT accuracy on Romanian verb conjugation, a computational linguistics evaluation of an LLM.", "model": "claude-haiku-5-5"}
++{"work_id": 4414790785, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.8, "rationale": "Studies autonomous behavior of LLM-based agents operating without tasks, with persistent memory and self-feedback.", "model": "claude-haiku-5-5"}
++{"work_id": 7214894185, "categories": ["llm", "computational_biology"], "primary": "llm", "confidence": 0.7, "rationale": "Retrieval and structure enhancement for oncology LLMs, with medical classification and QA benchmarks.", "model": "claude-haiku-5-5"}
++{"work_id": 7153669768, "categories": ["llm", "reinforcement_learning", "ai_safety"], "primary": "llm", "confidence": 0.7, "rationale": "Self-training generative reward models for LLM alignment using consistency-aware rewards.", "model": "claude-haiku-5-5"}
++{"work_id": 7155055855, "categories": ["llm", "deep_learning_architectures"], "primary": "llm", "confidence": 0.7, "rationale": "Knowledge distillation between LLMs with mismatched tokenizers, improving student LLM efficiency.", "model": "claude-haiku-5-5"}
++{"work_id": 7166569973, "categories": ["llm", "multimodal_ai", "ai_agents"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Multimodal LLM framework combining VLMs, audio models and RAG for customer service.", "model": "claude-haiku-5-5"}
++{"work_id": 7171730128, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Evaluation of tabular embedding similarity against human preference; weakly fits ML methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 7128492423, "categories": ["llm", "information_retrieval", "nlp"], "primary": "llm", "confidence": 0.8, "rationale": "RAG chatbot built on Mistral LLM for curriculum question answering; LLM-centred with retrieval component.", "model": "claude-haiku-5-5"}
++{"work_id": 4399996928, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Fine-tunes BERT for consumer sentiment and preference analysis, a classical text classification task rather than LLM research.", "model": "claude-haiku-5-5"}
++{"work_id": 4387819913, "categories": ["hci"], "primary": "hci", "confidence": 0.45, "rationale": "Qualitative study of pre-service teachers' experiences with ChatGPT in education; user-centred study framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7124885328, "categories": ["llm"], "primary": "llm", "confidence": 0.8, "rationale": "Survey of domain-specific LLM construction, continual pre-training and prompting.", "model": "claude-haiku-5-5"}
++{"work_id": 3185341429, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.8, "rationale": "Survey of prompt-based learning with pre-trained language models in NLP.", "model": "claude-haiku-5-5"}
++{"work_id": 7117486691, "categories": ["llm", "ai_agents"], "primary": "ai_agents", "confidence": 0.7, "rationale": "Covers LLM-based agentic AI, multi-agent collaboration, fine-tuning vs RAG.", "model": "claude-haiku-5-5"}
++{"work_id": 4417258277, "categories": ["llm", "deep_learning_architectures", "ml_theory"], "primary": "llm", "confidence": 0.6, "rationale": "Studies memorization vs generalization in capacity-limited pre-trained transformers, a learning-theory style analysis of LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4403025669, "categories": ["llm", "ai_safety"], "primary": "llm", "confidence": 0.55, "rationale": "Chapter on human evaluation and testing of LLMs, including ethics and cybersecurity use.", "model": "claude-haiku-5-5"}
++{"work_id": 4401044042, "categories": ["multimodal_ai", "llm", "speech_audio"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Aligns music audio encoder with frozen LLM for music captioning and QA; vision-language analogue for audio-text.", "model": "claude-haiku-5-5"}
++{"work_id": 7165003687, "categories": ["llm", "information_retrieval", "nlp"], "primary": "llm", "confidence": 0.7, "rationale": "Survey of LLM-powered question answering pipelines over text and tables, including query rewriting and planning.", "model": "claude-haiku-5-5"}
++{"work_id": 4416251417, "categories": ["ai_agents", "llm", "ai_safety", "sustainable_computing"], "primary": "llm", "confidence": 0.4, "rationale": "LLM-integrated decision-making toolkit for wildlife trade policy with optimisation module; LLM is central but the paper spans policy, curve fitting and bandit optimisation, so several categories are plausible but weak.", "model": "claude-haiku-5-5"}
++{"work_id": 7138971369, "categories": ["llm", "hci"], "primary": "llm", "confidence": 0.5, "rationale": "Case study of LLM-guided generation of Bebras computational thinking tasks with human collaboration; LLM use is the core object, with an education/human-AI collaboration angle.", "model": "claude-haiku-5-5"}
++{"work_id": 7125957422, "categories": ["llm", "multimodal_ai"], "primary": "llm", "confidence": 0.45, "rationale": "Multi-agent LLM architecture for natural-language interaction with smart meter data; primarily an LLM system application, arguably data-visualisation/database related.", "model": "claude-haiku-5-5"}
++{"work_id": 4414992583, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Systematic review of AI in K-12 language education from a pedagogical perspective; not CS research contribution in a core category.", "model": "claude-haiku-5-5"}
++{"work_id": 4387935330, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Legal and pedagogical discussion of AI in inclusive education; not computer-science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4391421531, "categories": ["llm"], "primary": "llm", "confidence": 0.9, "rationale": "Family of pre-trained and fine-tuned LLMs for creative writing with RAG and tool use and evaluation against GPT-4.", "model": "claude-haiku-5-5"}
++{"work_id": 4395022812, "categories": ["llm", "ml_theory"], "primary": "llm", "confidence": 0.7, "rationale": "Evaluation methodology for in-context learning of language models with a statistical baseline; LLM evaluation is primary, with methodological statistics.", "model": "claude-haiku-5-5"}
++{"work_id": 4414463094, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.75, "rationale": "Arabic dialect classification comparing RNNs, transformers and LLMs; classical NLP text classification framing dominates.", "model": "claude-haiku-5-5"}
++{"work_id": 7199029551, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Title suggests LLM interpretability/linguistic representations, but abstract is absent; conservatively a weak llm assignment is possible but the record is too sparse to classify.", "model": "claude-haiku-5-5"}
++{"work_id": 2904234310, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Constructs domain-specific sentiment dictionaries from reviews via representation learning; sentiment analysis is classical NLP.", "model": "claude-haiku-5-5"}
++{"work_id": 7140096463, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Network pharmacology and systems-level mechanism analysis of TCM prescriptions for RA, with EMR mining; biomedical computational methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4312852711, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Interpretable tabular neural model (Rules-TabNet) for ICU length-of-stay prediction; clinical application with a model-architecture contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7167456159, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Molecular docking and bioinformatic target prediction for antimicrobial compound.", "model": "claude-haiku-5-5"}
++{"work_id": 4296776208, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Local ancestry inference from genomic haplotype data, a genomics method.", "model": "claude-haiku-5-5"}
++{"work_id": 4386760409, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Network pharmacology and molecular docking of herbal formula for iron deficiency anemia.", "model": "claude-haiku-5-5"}
++{"work_id": 4402637119, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Radiomics and deep learning classification of breast masses in mammography and tomosynthesis images; medical image analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4410249860, "categories": ["computational_biology", "information_retrieval"], "primary": "computational_biology", "confidence": 0.4, "rationale": "QSAR and docking of phytochemicals with text mining for an educational platform; primarily life-science computation with an education framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7164761540, "categories": ["computer_vision", "computational_biology", "foundation_models"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Pathology foundation model aligning H&E slides with DNA methylation for cancer molecular prediction; multimodal histopathology and genomics.", "model": "claude-haiku-5-5"}
++{"work_id": 4394782468, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Digital twin simulation and integer programming for personalized neoantigen vaccine design.", "model": "claude-haiku-5-5"}
++{"work_id": 4400594688, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Bayesian multi-layer network model for spatial transcriptomics community connectivity.", "model": "claude-haiku-5-5"}
++{"work_id": 3127450329, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Study of clinician proficiency and training in EHR use; user-centred evaluation of interactive system, though health informatics framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3081811896, "categories": ["computational_biology", "generative_ai"], "primary": "computational_biology", "confidence": 0.3, "rationale": "Review of computer-aided synthesis planning for chemistry using AI; chemistry-oriented, weakly tied to deep learning generative methods. Low confidence and conservative assignment.", "model": "claude-haiku-5-5"}
++{"work_id": 4308207210, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Structure-based virtual screening and biological evaluation of drug candidates against a cancer target.", "model": "claude-haiku-5-5"}
++{"work_id": 3005160664, "categories": ["computational_biology", "computer_graphics"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Spatial visualization of tumor heterogeneity and cell interactions with agent-based modelling; biomedical informatics with visualization component.", "model": "claude-haiku-5-5"}
++{"work_id": 4391930238, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Review of machine learning for lung nodule detection in X-ray and CT images; medical image analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4389577171, "categories": ["ml_theory", "computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Continual learning for cuffless blood pressure estimation from PPG/ECG signals; biomedical signal ML application, learning approach is secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4389302773, "categories": ["computational_biology", "scientific_computing"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Free energy perturbation methodology for binding free energy in drug design; molecular simulation method.", "model": "claude-haiku-5-5"}
++{"work_id": 4411492298, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Open-source software tool for mass spectrometry data quality checking; computational tooling for analytical science.", "model": "claude-haiku-5-5"}
++{"work_id": 3163978416, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Mapping clinical data standards SDTM and FHIR for EHR real-world data; data integration and interoperability.", "model": "claude-haiku-5-5"}
++{"work_id": 4411977602, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Computational drug repurposing with docking and molecular dynamics targeting a disease protein.", "model": "claude-haiku-5-5"}
++{"work_id": 7135254915, "categories": [], "primary": null, "confidence": 0.9, "rationale": "In silico pharmacology/phytochemistry of plant extracts; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4411893160, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Automated segmentation and skeletonization of 2D microscopy images of dendritic spines; image analysis tool for neuroscience.", "model": "claude-haiku-5-5"}
++{"work_id": 7162037404, "categories": ["computational_biology", "ml_theory"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Inductive ontology-embedding method for disease-gene association prediction; bioinformatics with graph embedding methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4388076232, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Graph neural network method for lncRNA-disease association prediction; biological application with a GNN-style architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 3153063642, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Molecular complex characterization of receptor-ligand binding; computational structural biology and drug design.", "model": "claude-haiku-5-5"}
++{"work_id": 7221541424, "categories": ["computational_biology", "generative_ai"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Generative analogue design and docking for drug discovery; molecular generation is a component but the framing is drug discovery.", "model": "claude-haiku-5-5"}
++{"work_id": 3177565309, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Design and docking/MD of quinoline protease inhibitors for SARS-CoV-2; computational drug discovery.", "model": "claude-haiku-5-5"}
++{"work_id": 4318046762, "categories": ["multimodal_ai", "hci"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal annotation for pathology image classification and captioning with human-AI collaboration; vision-language plus interactive tool.", "model": "claude-haiku-5-5"}
++{"work_id": 7221738702, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Systems biology network data mining for cross-species drug target discovery.", "model": "claude-haiku-5-5"}
++{"work_id": 4302425986, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Network pharmacology analysis of coix seed for colorectal cancer; bioinformatics-style biomedical study.", "model": "claude-haiku-5-5"}
++{"work_id": 3174432239, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Curated database of reconstructed patient-level data from oncology trials; biomedical/clinical data methods, borderline with statistics and database work.", "model": "claude-haiku-5-5"}
++{"work_id": 7160950788, "categories": ["ml_theory", "deep_learning_architectures", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Graph neural network (QT-Net) and evaluation protocol for atomic-scale molecular property prediction; chemistry application with ML methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4405278807, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Conference proceedings front matter on general science; not clearly computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4220748066, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Network pharmacology and bioinformatics analysis of gene expression data.", "model": "claude-haiku-5-5"}
++{"work_id": 4411249474, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.75, "rationale": "Structure-based virtual screening and molecular dynamics for drug discovery.", "model": "claude-haiku-5-5"}
++{"work_id": 2783094078, "categories": ["computer_networks", "computational_biology"], "primary": "computer_networks", "confidence": 0.5, "rationale": "DNA-based storage via bacterial nanonetworks framed as networking with simulations and wet-lab validation; borderline between networking and synthetic biology.", "model": "claude-haiku-5-5"}
++{"work_id": 7135069038, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Survey of healthcare worker perceptions of digital pathology and AI; user attitudes study, though domain is medical informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 3173374233, "categories": ["generative_ai", "computational_biology"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Molecular graph generation and optimization with autoregressive learned distribution and Metropolis sampling for de novo drug design.", "model": "claude-haiku-5-5"}
++{"work_id": 2985643449, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Ligand-based pharmacophore modelling and biological evaluation for enzyme inhibitor discovery.", "model": "claude-haiku-5-5"}
++{"work_id": 7211895300, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "In silico molecular docking and ADMET study of plant compounds against bacterial gyrase.", "model": "claude-haiku-5-5"}
++{"work_id": 4402911702, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Statistical methodology for sparse multivariate additive regression with correlated errors; a statistics/ML methods paper applied to biology.", "model": "claude-haiku-5-5"}
++{"work_id": 4295929866, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "DeepLabv3 segmentation of vessels and tissue in endoscopic images and video, a medical image analysis task.", "model": "claude-haiku-5-5"}
++{"work_id": 3159313223, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Molecular docking and ADME analysis of phytochemicals against cytokine targets.", "model": "claude-haiku-5-5"}
++{"work_id": 4378771166, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Drug repurposing via molecular docking and QSAR regression against a viral protease.", "model": "claude-haiku-5-5"}
++{"work_id": 3110762405, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Deep neural networks predicting protein hydration properties for structure-activity and binding studies.", "model": "claude-haiku-5-5"}
++{"work_id": 4412915491, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Clinical decision support system for COVID-19 using health data; biomedical informatics, though weakly framed.", "model": "claude-haiku-5-5"}
++{"work_id": 4393899952, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Algorithm-based image analysis for counting mitochondria in immunohistochemistry sections, biomedical informatics method.", "model": "claude-haiku-5-5"}
++{"work_id": 4367393431, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.75, "rationale": "Ensemble learning for predicting disease-related miRNAs from bioinformatics databases.", "model": "claude-haiku-5-5"}
++{"work_id": 3020560080, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.4, "rationale": "Review of AI in breast medical imaging; primarily medical imaging analysis, though it is a review rather than a method.", "model": "claude-haiku-5-5"}
++{"work_id": 3159857369, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Comprehensive treatment of regularization methodology for machine learning on life science data.", "model": "claude-haiku-5-5"}
++{"work_id": 2786392319, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Essay on AI and climate change in general terms; no clear CS research contribution category dominates.", "model": "claude-haiku-5-5"}
++{"work_id": 3093665008, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.4, "rationale": "Literature review on managing uncertainty in AI-enabled decision making; closest fit is governance/reliability of AI, weakly classified.", "model": "claude-haiku-5-5"}
++{"work_id": 4387951271, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.3, "rationale": "HVAC building energy simulation with TRNSYS; building energy rather than computing footprint, so low fit and possibly not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4386212618, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.8, "rationale": "Estimates carbon emissions of website data transfer and proposes eco-friendly web design practices.", "model": "claude-haiku-5-5"}
++{"work_id": 7168018191, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Media and cultural studies qualitative review; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4410538484, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Fuzzy multi-criteria decision analysis of energy sources; operations research/energy domain rather than CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4313647803, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Consumer behavior study of online reviews in e-commerce; information systems marketing, not a CS ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4404238308, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Survey of solar power forecasting methods; energy domain with ML methods, weak fit to any CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4412066060, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.85, "rationale": "Testbed measuring energy and CO2 of data centre throughput computing and AI inference services.", "model": "claude-haiku-5-5"}
++{"work_id": 4403003621, "categories": ["hci", "sustainable_computing"], "primary": "hci", "confidence": 0.75, "rationale": "Focus-group user study on IoT repair and software obsolescence; HCI with sustainability framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3154667522, "categories": ["cloud_computing", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Green IT and cloud LMS for sustainable learning environment; cloud service design plus green computing framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7215545478, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Community service training on e-learning for teachers; education rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3111008797, "categories": ["deep_learning_architectures", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.5, "rationale": "LSTM occupancy prediction used for HVAC energy savings; energy management is the goal, LSTM is the method.", "model": "claude-haiku-5-5"}
++{"work_id": 4409494225, "categories": ["deep_learning_architectures", "ai_accelerators", "cybersecurity"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "NAS for adversarially robust approximate DNNs on accelerators; mixes architecture search, robustness and hardware efficiency.", "model": "claude-haiku-5-5"}
++{"work_id": 4308155833, "categories": ["computer_vision"], "primary": null, "confidence": 0.3, "rationale": "Remote-sensing reconstruction of CO2 via a neural network; primarily Earth observation application, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7034426089, "categories": ["software_engineering", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Model-based analysis of software energy consumption using LTS; green software engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 4400064539, "categories": ["sustainable_computing", "nlp"], "primary": "sustainable_computing", "confidence": 0.5, "rationale": "Energy-optimized loss for transformer selection in semantic communication; energy efficiency is central.", "model": "claude-haiku-5-5"}
++{"work_id": 4387983484, "categories": ["software_engineering", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Empirical study of energy consumption evolution in Kotlin Android apps across releases.", "model": "claude-haiku-5-5"}
++{"work_id": 3080194848, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Soil respiration prediction from spectral indices; agronomy/remote sensing rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4406505865, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Survey-based organizational readiness study in higher education; information systems social science rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7141737583, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Business and management book on AI strategy; not CS research contribution per se.", "model": "claude-haiku-5-5"}
++{"work_id": 6925471442, "categories": ["sustainable_computing", "software_engineering"], "primary": "sustainable_computing", "confidence": 0.8, "rationale": "Teaching green coding and energy-efficient software development course; sustainable computing with software engineering education framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7135017252, "categories": ["reinforcement_learning", "llm", "sustainable_computing"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "LLM-embedded multi-agent RL for low-carbon energy scheduling; main method is MARL with LLM-generated reward functions.", "model": "claude-haiku-5-5"}
++{"work_id": 4408774499, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Econometric ARDL study of CO2 emissions and AI innovation; economics rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4309156398, "categories": ["sustainable_computing", "distributed_computing"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Data center server power estimation and management to reduce power consumption.", "model": "claude-haiku-5-5"}
++{"work_id": 2996084711, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.65, "rationale": "Review of smartphone power consumption and energy management; energy footprint of computing systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7220836233, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Education management study on madrasahs; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4409426101, "categories": ["sustainable_computing", "distributed_computing", "cloud_computing"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Energy and workload management for geo-distributed data centers with renewables, using Stackelberg games to cut energy costs and emissions.", "model": "claude-haiku-5-5"}
++{"work_id": 3001058942, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Library science introduction on technical reports; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2948997629, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Critical systems analysis of a smart city ICT project with social and stakeholder focus; closest to HCI/social computing, sustainability framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3025387569, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Essay on digital ecology and socio-ecological transition; not a clear CS research contribution, so no category assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 4416214949, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Power-system electricity-carbon optimization for electrolytic aluminum load; energy/power engineering rather than a CS ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4386108360, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Sizing renewable and storage infrastructure for data centers to reduce carbon footprint; concerns the energy footprint of computing systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2905015752, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Compares CPU and GPU performance and power on smartphones for image processing; closest fit is hardware performance, though it is also computer vision applied to medical images.", "model": "claude-haiku-5-5"}
++{"work_id": 2950028967, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Demand-side carbon footprint allocation in power systems; power-systems and energy economics rather than a CS category, though it touches on smart grids.", "model": "claude-haiku-5-5"}
++{"work_id": 4392157044, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Economics panel study of ICT, gender and CO2 emissions; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4207059814, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Energy profiling tool for Android apps supports green software engineering; could also be software_engineering, but energy footprint is the central focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4415684933, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Economics/finance policy analysis using DID on listed firms; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 2990816793, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Survey of e-waste from electronic devices and its environmental implications for ICT researchers; sustainability of computing hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4323665427, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Urban green space and building carbon emissions using remote sensing and spatial analysis; environmental science rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4399765326, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Photovoltaic energy simulation study using PVsyst; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7124145190, "categories": ["sustainable_computing", "llm"], "primary": "sustainable_computing", "confidence": 0.8, "rationale": "Energy efficiency metric for AI servers under LLM workloads; main contribution is green computing measurement.", "model": "claude-haiku-5-5"}
++{"work_id": 4406559909, "categories": ["ai_agents"], "primary": null, "confidence": 0.2, "rationale": "RBF neural network MPPT controller for PV; control application, weakly fits no ontology category well, so minimal assignment with low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7117113370, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "MaxEnt habitat suitability modelling for dolphin conservation is ecological rather than CS; closest fit is computational biology methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4390978249, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Provincial carbon emission forecasting with LEAP model; energy/policy analysis, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3036320021, "categories": ["distributed_computing", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Energy-aware meta-heuristic scheduling in heterogeneous virtual cloud; green scheduling for cloud resources.", "model": "claude-haiku-5-5"}
++{"work_id": 4361987249, "categories": [], "primary": null, "confidence": 0.6, "rationale": "ANFIS-GA prediction of PV electrical output; applied energy forecasting rather than CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4399284355, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Deep learning forecasting of agricultural carbon emissions; environmental science application, not CS core category.", "model": "claude-haiku-5-5"}
++{"work_id": 4390621478, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Quiz-based challenge raising awareness of digital technology impacts; user study and education design.", "model": "claude-haiku-5-5"}
++{"work_id": 2990724040, "categories": ["sustainable_computing", "computer_architecture"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Data center UPS consolidation to reduce power losses; energy efficiency of computing infrastructure.", "model": "claude-haiku-5-5"}
++{"work_id": 7163329089, "categories": ["edge_computing"], "primary": null, "confidence": 0.1, "rationale": "Educational evaluation of an LMS for blended learning; not clearly a CS category, so essentially non-CS education research with no applicable category.", "model": "claude-haiku-5-5"}
++{"work_id": 4367163963, "categories": ["quantum_computing", "deep_learning_architectures"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Shot optimization for training hybrid quantum-classical QML models, accelerating quantum machine learning training.", "model": "claude-haiku-5-5"}
++{"work_id": 4225771907, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Quantum communication with space-encoded quantum states, a quantum networking/communication topic.", "model": "claude-haiku-5-5"}
++{"work_id": 4415252535, "categories": ["quantum_computing", "federated_learning"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Distributed quantum ML with knowledge distillation across heterogeneous clients; quantum models with federated-style aggregation.", "model": "claude-haiku-5-5"}
++{"work_id": 7116076303, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.5, "rationale": "Quantum-correlation-based radiometric calibration of photodiodes for optical quantum computing; primarily quantum optics/metrology with quantum information framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3206680160, "categories": ["quantum_computing", "applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Quantum key distribution protocol analysis using thermal states, a cryptographic protocol design and security analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4417308968, "categories": ["quantum_computing", "distributed_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Circuit partitioning and cutting with hybrid classical-quantum execution for NISQ devices.", "model": "claude-haiku-5-5"}
++{"work_id": 7131679394, "categories": ["quantum_computing", "computer_architecture"], "primary": "quantum_computing", "confidence": 0.5, "rationale": "Superconducting parametric amplifier for quantum readout based on granular aluminum; quantum hardware device physics.", "model": "claude-haiku-5-5"}
++{"work_id": 4412565510, "categories": ["quantum_computing", "computer_architecture"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Chip-based 3D ion traps for scalable qubit hardware platforms.", "model": "claude-haiku-5-5"}
++{"work_id": 2896762452, "categories": ["quantum_computing", "scientific_computing"], "primary": "quantum_computing", "confidence": 0.5, "rationale": "First-principles DFT study of electric field noise from adsorbates causing motional heating in ion-trap quantum computers; qubit hardware physics with computational methods.", "model": "claude-haiku-5-5"}
++{"work_id": 7125762882, "categories": ["quantum_computing", "applied_cryptography"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "4D QKD system over multicore fibre; quantum key distribution is a quantum communication topic, with cryptographic protocol flavour.", "model": "claude-haiku-5-5"}
++{"work_id": 2964632298, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Reversible logic circuit design (code converter) with quantum cost metrics; digital circuit design, though reversible computing is quantum-adjacent.", "model": "claude-haiku-5-5"}
++{"work_id": 4296884690, "categories": ["applied_cryptography", "quantum_computing"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Security of QKD systems against laser damage attacks, protecting optical components of key distribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4415968028, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Quantum metrology with nonlinear interactions and noise; quantum sensing/information theory framing.", "model": "claude-haiku-5-5"}
++{"work_id": 2902876714, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Thesis on coherent control and sideband cooling of trapped ions for quantum control.", "model": "claude-haiku-5-5"}
++{"work_id": 2902785919, "categories": ["quantum_computing", "scientific_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Hybrid NISQ algorithm simulating scalar quantum field theory and benchmarking quantum hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 7163287303, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Experimental quantum game implementation on NISQ hardware with noise analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 2783787559, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Speculative analogy between spin glass magnetism and quantum notions for molecular networks; not clearly a CS contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7161271332, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Multi-copy entanglement detection with experiments on IBM quantum processors.", "model": "claude-haiku-5-5"}
++{"work_id": 3112102734, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Technical perspective on reducing error-correction cost of quantum computation.", "model": "claude-haiku-5-5"}
++{"work_id": 4200635042, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Spin squeezing and non-Gaussian entangled states for quantum metrology; quantum physics with information/metrology framing, closest fit is quantum_computing, though it is largely quantum physics.", "model": "claude-haiku-5-5"}
++{"work_id": 4312768793, "categories": ["applied_cryptography", "computer_networks"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Review of quantum key distribution for securing visible light communication networks.", "model": "claude-haiku-5-5"}
++{"work_id": 2964298217, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Quantum optics interferometry and phase sensitivity; physics rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4411232250, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.35, "rationale": "Unambiguous state discrimination of coherent states for future quantum communication networks; borderline between quantum communication and physics.", "model": "claude-haiku-5-5"}
++{"work_id": 4385967641, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Mean-field analysis of IBM's utility experiment simulating kicked Ising dynamics on a quantum computer.", "model": "claude-haiku-5-5"}
++{"work_id": 4411225084, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Superconducting resonator materials and fabrication for qubits; hardware physics, arguably quantum hardware but not clearly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3161468414, "categories": ["quantum_computing", "computational_biology"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "VQE-based quantum algorithm for bosonic simulations with resource reduction, applied to molecular vibrational properties.", "model": "claude-haiku-5-5"}
++{"work_id": 4400861003, "categories": ["quantum_computing", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "GPU-accelerated tensor network contraction for classical simulation of quantum circuits; high-performance numerical computing.", "model": "claude-haiku-5-5"}
++{"work_id": 4417073300, "categories": ["quantum_computing", "cybersecurity"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Software for evaluating cloud quantum computers via random bit sequence generation and NIST tests, related to quantum random number generation for security.", "model": "claude-haiku-5-5"}
++{"work_id": 4416775576, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Mathematical analysis of entanglement witnesses for detecting Bell states, a quantum information topic.", "model": "claude-haiku-5-5"}
++{"work_id": 3035347578, "categories": ["quantum_computing", "deep_learning_architectures"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Quantum reservoir computing for temporal learning; quantum systems framing primary, with reservoir/learning component.", "model": "claude-haiku-5-5"}
++{"work_id": 3137888032, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Thesis on quantum reference frames and subsystem decompositions for quantum information tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 4391889616, "categories": ["quantum_computing", "applied_cryptography"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Twin-field QKD protocol with security proof; quantum key distribution is quantum communication.", "model": "claude-haiku-5-5"}
++{"work_id": 7148276906, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Quantum state preparation algorithm combining VQAs and adiabatic preparation.", "model": "claude-haiku-5-5"}
++{"work_id": 2995678717, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Continuous dynamical decoupling to protect qudit quantum gates from noise.", "model": "claude-haiku-5-5"}
++{"work_id": 4387681983, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Bayesian optimization of QAOA parameters on quantum circuits.", "model": "claude-haiku-5-5"}
++{"work_id": 4387559660, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Hybrid classical-quantum state loading of probability distributions on IBM quantum hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4400297779, "categories": ["quantum_computing", "ml_theory"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Quantum machine learning with entangled data and no-free-lunch analysis of prediction error.", "model": "claude-haiku-5-5"}
++{"work_id": 2755062554, "categories": ["quantum_computing", "computer_architecture"], "primary": "quantum_computing", "confidence": 0.4, "rationale": "Single-photon temporal-spectral characterization for quantum optical technologies; experimental quantum photonics, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4404573991, "categories": ["quantum_computing", "computational_complexity"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Entanglement robustness in quantum networks related to graph connectivity; quantum information theory with combinatorial asymptotics.", "model": "claude-haiku-5-5"}
++{"work_id": 4402268320, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Topological features of entangled states relevant to quantum technologies and noise robustness; quantum information framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4410302895, "categories": ["quantum_computing", "computational_complexity"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Learning quantum processes with query complexity lower bounds and quantum hardware security application.", "model": "claude-haiku-5-5"}
++{"work_id": 4221156287, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Quantum error correction in the 3D surface code with non-Pauli errors and magic state preparation.", "model": "claude-haiku-5-5"}
++{"work_id": 4297796489, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.75, "rationale": "Photon-pair high-dimensional entanglement generation for quantum information; photonic quantum technology.", "model": "claude-haiku-5-5"}
++{"work_id": 7110562677, "categories": ["quantum_computing", "scientific_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Tensor network simulation of quantum circuits with VQE implementation and application to vehicle routing.", "model": "claude-haiku-5-5"}
++{"work_id": 4389163069, "categories": ["quantum_computing", "applied_cryptography"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Semi-quantum random number generation with security proof; quantum cryptography.", "model": "claude-haiku-5-5"}
++{"work_id": 2996620206, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Generation of photonic linear cluster states from a single photon emitter for one-way quantum computing.", "model": "claude-haiku-5-5"}
++{"work_id": 4393118756, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Quantum algorithms to probe quantum geometric tensor on IBM quantum computers, variational and imaginary-time methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4401437211, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Educational simulation of decoherence in superconducting qubits with Lindblad master equation.", "model": "claude-haiku-5-5"}
++{"work_id": 7127900880, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Improved Rodeo algorithm time sampling for spectrum computation and state preparation.", "model": "claude-haiku-5-5"}
++{"work_id": 4390533496, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Feature stores are data management infrastructure for ML pipelines; closest fit is database systems, though it is ML-systems oriented.", "model": "claude-haiku-5-5"}
++{"work_id": 7162347486, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Flash translation layer address mapping in SSD hardware/firmware is storage architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 4394871762, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.9, "rationale": "OS compatibility layers for applications on new operating systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3023238978, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Cloud resource provisioning and limit specification to reduce wastage.", "model": "claude-haiku-5-5"}
++{"work_id": 2964142687, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.55, "rationale": "SSD data layout and read performance across parallel units is storage hardware/system design; could also be operating_systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4400484472, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Augmented reality mobile app for visitor guidance in museums, with user-experience focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4391306053, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Inter-cloud data transfer performance with adaptive buffering between compute and storage clouds.", "model": "claude-haiku-5-5"}
++{"work_id": 4292057429, "categories": ["information_retrieval", "computer_networks"], "primary": "computer_networks", "confidence": 0.4, "rationale": "Private information retrieval over distributed storage codes; closer to coding/storage networking, with privacy aspect but not privacy_preserving_computation primarily; low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 3027986714, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.75, "rationale": "New operating system architecture blending distributed OS and unikernel ideas.", "model": "claude-haiku-5-5"}
++{"work_id": 4254652521, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.35, "rationale": "Optical storage hardware and servo systems; hardware-oriented but not clearly within core categories.", "model": "claude-haiku-5-5"}
++{"work_id": 7204340147, "categories": ["cloud_computing", "cybersecurity", "operating_systems"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Confidential VMs with SEV memory encryption against malicious hypervisors in cloud nested virtualization; security of VM isolation is the core contribution, with cloud and OS-level virtualization aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 4225872280, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Practical book chapter on Ansible modules for server administration; not a research contribution clearly in a CS category beyond general systems tooling, so no category assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 3189882660, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Commercial price list for VMware licenses; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4224674551, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Statistical distribution modelling of COVID-19 waves; statistics/epidemiology rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4406461284, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Experimental characterization of read disturb in 3D NAND flash memory hardware, relevant to memory system design and reliability.", "model": "claude-haiku-5-5"}
++{"work_id": 4293193165, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Digital predistortion for RF power amplifier linearization is signal-processing hardware; closest fit is computer_architecture/embedded hardware, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 3033055137, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.8, "rationale": "File system design with many-to-one file-to-metadata mapping, a storage systems contribution in OS territory.", "model": "claude-haiku-5-5"}
++{"work_id": 3167972790, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.6, "rationale": "Describes NTFS hidden features (ADS, hard links, junctions) for digital media collectors; file system topic, though practical and lightweight.", "model": "claude-haiku-5-5"}
++{"work_id": 7117899816, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Theoretical socio-political framework about post-crisis regimes; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3095048882, "categories": ["database_systems", "operating_systems"], "primary": "operating_systems", "confidence": 0.6, "rationale": "I/O deduplication for primary storage systems in the cloud, implemented in the Linux kernel; storage system performance with data management aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 4380875641, "categories": ["ml_theory", "database_systems"], "primary": "ml_theory", "confidence": 0.3, "rationale": "NILM appliance classification with feature selection fusion and Bayesian optimization; primarily a signal/ML classification application, weakly fits ML methodology. Could also be speech_audio-free energy application; low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4213090496, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.85, "rationale": "IO control for containers in datacenter block storage, kernel-level resource isolation.", "model": "claude-haiku-5-5"}
++{"work_id": 2902405086, "categories": ["operating_systems", "database_systems"], "primary": "database_systems", "confidence": 0.7, "rationale": "Full-path indexed file system built on write-optimized dictionaries (BetrFS); storage and key-value dictionary design.", "model": "claude-haiku-5-5"}
++{"work_id": 4226300713, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Azure Virtual Desktop storage configuration (FSLogix), a cloud desktop service book chapter.", "model": "claude-haiku-5-5"}
++{"work_id": 4401508056, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.6, "rationale": "ZNS SSD firmware and I/O stack write ordering; storage device systems, could also be computer_architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 3173583453, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.6, "rationale": "Performance comparison of Windows 10 and Linux Mint operating systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2901264075, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.6, "rationale": "SMR disk management and cleaning scheme for storage, file/storage system level.", "model": "claude-haiku-5-5"}
++{"work_id": 2973989512, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Tiered data storage placement algorithm for files using Kohonen network; storage management, uses neural network tool.", "model": "claude-haiku-5-5"}
++{"work_id": 4386952147, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Caching for low latency in distributed erasure-coded storage systems with sampling-based cache partitioning.", "model": "claude-haiku-5-5"}
++{"work_id": 4387067117, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Practical iOS app persistence options guide; mobile app development topic, weak fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3181207510, "categories": ["software_engineering", "hci"], "primary": "hci", "confidence": 0.5, "rationale": "Mobile app for classroom use by teachers; interactive system design for education, with app development aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 2894822167, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.3, "rationale": "Benchmarking CernVM-FS file distribution system; a file-system/software-distribution study, low information.", "model": "claude-haiku-5-5"}
++{"work_id": 4401070015, "categories": ["ml_theory", "computer_vision"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Statistical method (IVA with bootstrap hypothesis testing) for relationship structure; applied to fMRI, so neuroimaging application is secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4392711967, "categories": ["database_systems", "computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Provisioning tool for computational storage drive arrays with erasure coding offloaded to CSDs; storage hardware system design.", "model": "claude-haiku-5-5"}
++{"work_id": 3097083941, "categories": ["distributed_computing", "software_engineering"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Scalable online code execution system architecture; a distributed deployable service, with education and competitive programming uses.", "model": "claude-haiku-5-5"}
++{"work_id": 3212917515, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Experiment on learner engagement and self-learning with multi-abstraction RTOS tools in engineering education.", "model": "claude-haiku-5-5"}
++{"work_id": 7137550668, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Lightning talk slide timestamps for summer students; not substantive research content.", "model": "claude-haiku-5-5"}
++{"work_id": 7213061804, "categories": ["deep_learning_architectures", "llm"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "MoE model inference offloading with sequence-based expert prediction; efficient inference of neural architectures, LLM-adjacent.", "model": "claude-haiku-5-5"}
++{"work_id": 4408213108, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Abstract is a quote with no computer science research content apparent.", "model": "claude-haiku-5-5"}
++{"work_id": 3170878721, "categories": ["operating_systems", "computer_architecture"], "primary": "operating_systems", "confidence": 0.5, "rationale": "Lightweight hypervisor architecture for mixed-criticality IoT systems; virtualisation and scheduling in system software.", "model": "claude-haiku-5-5"}
++{"work_id": 3008989608, "categories": ["database_systems", "distributed_computing"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Improves data movement to object storage via client-server cooperative merging and splitting; storage and data transfer systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4224080142, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.9, "rationale": "CPU round-robin scheduling algorithm with dynamic time quantum, a classic OS scheduling contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3193637602, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Comparative study of mobile piano-learning apps in education, an interactive-system and user-facing evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 2981636369, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Framework for adapting learning material to student skills in e-learning; weakly aligned with interactive systems and learners, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4394586023, "categories": ["cloud_computing", "distributed_computing"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Trace analysis and optimization of container image transfer and storage in container registries for cloud environments.", "model": "claude-haiku-5-5"}
++{"work_id": 4408867919, "categories": ["database_systems", "operating_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Defragmentation for deduplicated backup storage; storage management of backup data.", "model": "claude-haiku-5-5"}
++{"work_id": 7137478203, "categories": ["distributed_computing", "cloud_computing"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Workshop talk on CVMFS software distribution infrastructure and storage backends at CERN.", "model": "claude-haiku-5-5"}
++{"work_id": 4200058389, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Bayesian and classical estimation for a process capability index is statistics rather than CS; conservative, likely not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3022441546, "categories": ["nlp"], "primary": null, "confidence": 0.1, "rationale": "Data science curriculum for K-12 education about AI; not clearly a CS research contribution in any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 3113396090, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Nonlinear plasmonics in gold nanostructures is physics/optics, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3082840770, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "Reconfigurable FPGA architecture design with reduced area and SRAM; hardware architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 3112282291, "categories": ["computer_architecture", "ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "In-storage computing hardware accelerating hyperdimensional computing encoding and training with FPGA and flash circuits.", "model": "claude-haiku-5-5"}
++{"work_id": 4402895027, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.7, "rationale": "In-field testing of transition faults for chip reliability; hardware testing and reliability.", "model": "claude-haiku-5-5"}
++{"work_id": 3142919340, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "Clock distribution design for RSFQ superconducting circuits; electronic design automation.", "model": "claude-haiku-5-5"}
++{"work_id": 3038664825, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Hot-carrier degradation of power LDMOS transistors; device reliability physics rather than computing, closest fit is hardware reliability.", "model": "claude-haiku-5-5"}
++{"work_id": 4403361878, "categories": ["computer_architecture", "computer_graphics"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "FPGA-based sprite drawing hardware for mobile terminals; hardware design with graphics application.", "model": "claude-haiku-5-5"}
++{"work_id": 4411019384, "categories": ["cybersecurity", "ml_theory"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Deception attack on DOA estimation algorithms; adversarial security framing with signal processing optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 4382052984, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "Survey of CPU instruction pipelines for RISC and CISC architectures.", "model": "claude-haiku-5-5"}
++{"work_id": 4414630583, "categories": ["ai_accelerators", "distributed_computing", "information_retrieval"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "CPU-NPU heterogeneous offloading for vector embedding in RAG inference serving systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4407949197, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Fault diagnosis strategy for multiprocessor systems under PMC and MM models.", "model": "claude-haiku-5-5"}
++{"work_id": 3040997825, "categories": ["computer_architecture", "ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "Mapping 3D CNNs onto a vector DSP with Winograd optimization is hardware-aware ML acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 4383502097, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.7, "rationale": "QCA circuit design of XOR gates and code converters is digital logic/nanoscale circuit design.", "model": "claude-haiku-5-5"}
++{"work_id": 4400224459, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "Minimal-area scalable RISC-V processor core implemented on FPGAs is processor microarchitecture design.", "model": "claude-haiku-5-5"}
++{"work_id": 2809334893, "categories": ["computer_architecture", "computer_networks"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Current sensor SoC with integrated compressive sensing encoder is mixed-signal hardware design; framed as sensing SoC rather than networking.", "model": "claude-haiku-5-5"}
++{"work_id": 7206165071, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.75, "rationale": "Multi-gate transistor-based adder/subtractor logic design for low-power arithmetic circuits.", "model": "claude-haiku-5-5"}
++{"work_id": 2942318936, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Model-based testing and quality assurance for component-based embedded applications.", "model": "claude-haiku-5-5"}
++{"work_id": 2966831047, "categories": ["distributed_computing", "deep_learning_architectures"], "primary": "distributed_computing", "confidence": 0.75, "rationale": "Hybrid data and model parallelism for distributed multi-GPU deep learning training systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3158474254, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Side-channel attack extracting BNN architecture and secret weights from hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 7160405187, "categories": ["formal_verification", "computational_complexity"], "primary": "formal_verification", "confidence": 0.4, "rationale": "Reachability computation of synchronous automata using ROBDDs, closer to automated reasoning and circuit verification/testing.", "model": "claude-haiku-5-5"}
++{"work_id": 7133560921, "categories": ["computer_architecture", "operating_systems", "cybersecurity"], "primary": "computer_architecture", "confidence": 0.75, "rationale": "Hardware memory controller design for CXL memory with reliability and security mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 2946432794, "categories": ["ai_accelerators", "ml_theory"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Brain-inspired hyperdimensional clustering with in-memory processing; primarily an unsupervised learning algorithm with hardware-acceleration framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3206763996, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.85, "rationale": "Approximate 4-2 compressors for multiplier circuit design, VLSI implementation in 28nm.", "model": "claude-haiku-5-5"}
++{"work_id": 4410057699, "categories": ["algorithms", "distributed_computing"], "primary": "algorithms", "confidence": 0.6, "rationale": "Parallel GPU-accelerated multilevel graph partitioning algorithm for CAD; algorithmic design with parallel systems aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 3142700669, "categories": ["algorithms", "programming_languages"], "primary": "algorithms", "confidence": 0.5, "rationale": "Graph-theoretic products for CSP process scheduling/automation; combinatorial analysis of process algebra communication.", "model": "claude-haiku-5-5"}
++{"work_id": 4399426830, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Data engineering pipeline (Medallion architecture) analyzing CPU performance metrics.", "model": "claude-haiku-5-5"}
++{"work_id": 4386926746, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Temperature-based covert channel in FPGA systems, a hardware security attack study.", "model": "claude-haiku-5-5"}
++{"work_id": 3185457619, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.9, "rationale": "FPGA coprocessor design for CNN acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 2886646777, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Time synchronization between FlexRay and Ethernet automotive networks (IEEE 802.1AS).", "model": "claude-haiku-5-5"}
++{"work_id": 3082922165, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.35, "rationale": "Real-time schedulability analysis for multiframe task systems; closest fit among ontology is systems scheduling, though it is real-time systems rather than distributed.", "model": "claude-haiku-5-5"}
++{"work_id": 4386764847, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Software-controlled covert channel attacks based on mutual exclusion mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 2888777304, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Spin-orbit torque charge-to-spin conversion in topological insulator bilayers for SOT-MRAM; materials/device physics with memory-hardware motivation, borderline for the ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 3147171055, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Transistor device and analog circuit design (LTFET vs MOSFET) for VLSI; fits low-power/VLSI device-circuit design.", "model": "claude-haiku-5-5"}
++{"work_id": 4367663628, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.75, "rationale": "Cache hierarchy evaluation with MRAM/SRAM and prefetching on RISC-V; memory systems architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 2784201766, "categories": ["cybersecurity", "computer_architecture"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Memristor crossbar physical unclonable functions for hardware security; abstract is truncated.", "model": "claude-haiku-5-5"}
++{"work_id": 4416176040, "categories": ["llm", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.8, "rationale": "Carbon footprint modeling of LLM training under neural scaling laws; sustainability is the main contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4415481175, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Integrated circuit energy harvesting platform for IoT with CMOS prototype; circuit design with hardware focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3162223152, "categories": ["cybersecurity", "computer_architecture"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "FPGA RO-PUF quality improvement for device authentication; hardware security primitive.", "model": "claude-haiku-5-5"}
++{"work_id": 7170236789, "categories": ["scientific_computing", "distributed_computing"], "primary": "scientific_computing", "confidence": 0.8, "rationale": "Iterative numerical methods for PDEs with OpenMP parallelisation on multi-core.", "model": "claude-haiku-5-5"}
++{"work_id": 4416146873, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Reproducibility report on GPU resilience characterization; GPU reliability is hardware-focused, though it is a meta report.", "model": "claude-haiku-5-5"}
++{"work_id": 3010845351, "categories": ["distributed_computing", "algorithms"], "primary": "distributed_computing", "confidence": 0.7, "rationale": "Parallel HPC implementation of DBSCAN clustering on modular supercomputer architecture for exascale.", "model": "claude-haiku-5-5"}
++{"work_id": 2800869073, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Analog circuit synthesis and performance evaluation of circuit topologies; electronic design automation falls under computer_architecture (EDA/VLSI).", "model": "claude-haiku-5-5"}
++{"work_id": 4213103670, "categories": ["cybersecurity", "operating_systems"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Covert channel exploiting DVFS vulnerability in kernel/userspace; a side-channel security attack.", "model": "claude-haiku-5-5"}
++{"work_id": 4392928691, "categories": ["ai_accelerators", "computer_vision"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "FPGA accelerator design for CNN-based YOLO-V3 object detection.", "model": "claude-haiku-5-5"}
++{"work_id": 2889574478, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Memory access profiling for application-specific memory subsystem design in hardware architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 3023376085, "categories": ["distributed_computing", "programming_languages"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Programming BSP and Multi-BSP parallel algorithms; abstract is empty so judgment is from title only.", "model": "claude-haiku-5-5"}
++{"work_id": 4256423899, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Parametric fault detection in analog circuits for testability; hardware test engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 3199043212, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Test pattern reordering to reduce test power in VLSI circuits.", "model": "claude-haiku-5-5"}
++{"work_id": 3004719967, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Railway traffic safety and repair risk assessment; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2911751195, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "Open-source FPGA architecture and circuit design toolset for heterogeneous blocks.", "model": "claude-haiku-5-5"}
++{"work_id": 4312741045, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Compact device model of phase change memory for circuit simulation; device-level modelling for memory hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4411122680, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Simulation of tonal rail track circuit electronics; railway signalling, not clearly a listed CS category, weakly fits embedded/electronic design.", "model": "claude-haiku-5-5"}
++{"work_id": 4416959540, "categories": ["computational_complexity", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Certified homotopy tracking for polynomial systems with interval arithmetic and complexity bounds for numerical solvers.", "model": "claude-haiku-5-5"}
++{"work_id": 4415164777, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Data-driven predictive control with Hankel matrices and SVD; control theory rather than a clear CS category, weakly fits ml_theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4235658738, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Thomas algorithm for tridiagonal linear systems, numerical linear algebra.", "model": "claude-haiku-5-5"}
++{"work_id": 4387706174, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Continuous-time inertial dynamics for convex optimization; mathematical optimization analysis, closest to numerical methods.", "model": "claude-haiku-5-5"}
++{"work_id": 6940648869, "categories": [], "primary": null, "confidence": 0.4, "rationale": "LQR design for hybrid power plant load sharing; control engineering rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4224920451, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.45, "rationale": "Data-driven symbolic regression for Floquet propagator inverse problem in acoustics and material parameter estimation.", "model": "claude-haiku-5-5"}
++{"work_id": 4363675873, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "B-spline Galerkin numerical approximation of a time-fractional PDE.", "model": "claude-haiku-5-5"}
++{"work_id": 2969289037, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Python software package for accessing solar observatory data; scientific software tooling.", "model": "claude-haiku-5-5"}
++{"work_id": 4205760564, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Book chapter on linear algebra algorithms and implementations for numerical computing.", "model": "claude-haiku-5-5"}
++{"work_id": 6967178263, "categories": ["ml_theory", "computational_biology", "foundation_models", "scientific_computing"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Evaluates universal machine-learned interatomic potentials for biomolecular simulation; application to biomolecules is the core framing, with foundation-model-style evaluation of transferable potentials.", "model": "claude-haiku-5-5"}
++{"work_id": 3134502016, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Commentary on secondary analysis of datasets and research reproducibility in science; not a CS research contribution in a listed category.", "model": "claude-haiku-5-5"}
++{"work_id": 4404306456, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "Numerical virtual element method for magnetohydrodynamics with error estimates; discretisation of PDEs.", "model": "claude-haiku-5-5"}
++{"work_id": 7152139045, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Domain decomposition and multiscale numerical methods for urban flood PDE models.", "model": "claude-haiku-5-5"}
++{"work_id": 4327485924, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Coupling PyTorch with Firedrake for PDE-based machine learning and simulation frameworks.", "model": "claude-haiku-5-5"}
++{"work_id": 4406816254, "categories": ["scientific_computing"], "primary": null, "confidence": 0.3, "rationale": "Inventory and lateral transshipment simulation-optimisation is operations research, not clearly any listed CS category; weakly assigned to nothing conservative.", "model": "claude-haiku-5-5"}
++{"work_id": 4376654240, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Iterative numerical scheme for thermal radiative transfer equations with time integration.", "model": "claude-haiku-5-5"}
++{"work_id": 4206279532, "categories": ["nlp"], "primary": null, "confidence": 0.1, "rationale": "Neural network passenger flow prediction for railways; applied ML forecasting that does not clearly fit any listed category, so left unassigned.", "model": "claude-haiku-5-5"}
++{"work_id": 3199718208, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Numerical discretisation of stochastic thin-film equation with structure preservation.", "model": "claude-haiku-5-5"}
++{"work_id": 4313597919, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "High-order DGSEM with shock capturing for compressible fluid dynamics.", "model": "claude-haiku-5-5"}
++{"work_id": 3210446440, "categories": ["algorithms", "information_retrieval"], "primary": "algorithms", "confidence": 0.4, "rationale": "Bi-criteria optimization of public transport planning using local search and genetic algorithms with a machine-learned robustness oracle; primarily optimization/metaheuristics, with ML oracle as surrogate.", "model": "claude-haiku-5-5"}
++{"work_id": 4405077071, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.3, "rationale": "Electromagnetic tomography image reconstruction and parameter optimisation; imaging inverse problems, closest to image reconstruction in computer vision/engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 7155389303, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.8, "rationale": "Preconditioned conjugate gradient solvers for boundary element linear systems; numerical methods.", "model": "claude-haiku-5-5"}
++{"work_id": 4291941306, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Lie symmetry analysis of hyperbolic PDE model; applied mathematics rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3216320654, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Gradient-based adaptive optimal control with parameter estimation for unknown linear systems; control theory with learning-style estimation, borderline.", "model": "claude-haiku-5-5"}
++{"work_id": 3084714138, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Workflow importing omics metadata into data papers; data management and metadata publishing, borderline with scientific data.", "model": "claude-haiku-5-5"}
++{"work_id": 4408094557, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Nonlinear tyre dynamics PDE modelling and stability analysis; mechanical engineering/applied mathematics.", "model": "claude-haiku-5-5"}
++{"work_id": 4417347238, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Algebraic/optimization algorithm for tensor CP decomposition; numerical multilinear algebra, could also fit algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 2963395393, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Homogenization of elliptic PDE systems; pure mathematics.", "model": "claude-haiku-5-5"}
++{"work_id": 4382681213, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Iterative root-finding methods for polynomials convergence analysis; numerical analysis, mostly mathematics.", "model": "claude-haiku-5-5"}
++{"work_id": 7163167038, "categories": ["scientific_computing", "algorithms"], "primary": "scientific_computing", "confidence": 0.8, "rationale": "Nonlinear domain decomposition preconditioners for PDE solvers, Newton methods, applied in Code_Aster; numerical methods for scientific simulation.", "model": "claude-haiku-5-5"}
++{"work_id": 3082621528, "categories": ["computer_architecture"], "primary": null, "confidence": 0.2, "rationale": "Bio-impedance electrode design via FEM for skin measurement; biomedical engineering rather than CS, low confidence in any CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 3012277325, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Numerical linear algebra study of power method eigenvalue acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 4414593075, "categories": ["computer_vision", "deep_learning_architectures", "scientific_computing"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Neural operator for 3D photoacoustic tomography image reconstruction from sensor data.", "model": "claude-haiku-5-5"}
++{"work_id": 4392110244, "categories": ["scientific_computing"], "primary": null, "confidence": 0.3, "rationale": "Equivalent heat source modelling of AC arcs and fire risk; physics simulation with neural network and GA, weakly CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3157604584, "categories": ["scientific_computing"], "primary": null, "confidence": 0.25, "rationale": "Mathematical dynamical systems analysis of swarming on sphere; applied mathematics rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4394993168, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Epidemiological-style corruption dynamics model with optimal control; applied mathematics, not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 2935049866, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Model reduction for discrete-time systems in control theory; closest CS-adjacent category is weak.", "model": "claude-haiku-5-5"}
++{"work_id": 3159820714, "categories": ["scientific_computing"], "primary": null, "confidence": 0.4, "rationale": "LES combustion modelling of flashback in gas turbines; computational fluid dynamics rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3000672483, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Stochastic quantization path integral framing for constrained Ito processes and Kalman filter derivation; probabilistic modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 7168156340, "categories": ["deep_learning_architectures", "speech_audio"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Adaptive filtering (RLS) with tensor decomposition for echo cancellation; signal-processing algorithm closest to model design, with acoustic echo application.", "model": "claude-haiku-5-5"}
++{"work_id": 3088835008, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.3, "rationale": "Beidou-GPS dual-mode positioning system on mobile devices; satellite positioning and mobile communication equipment, loosely fitting networks or systems; low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4390667245, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Semi-analytical numerical method for population balance equations with convergence analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4322394327, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.4, "rationale": "Co-simulation of power system and ICT networks for smart grids; overview of simulation interfacing, closest to networking/distributed simulation.", "model": "claude-haiku-5-5"}
++{"work_id": 2963298714, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Sparse matrix storage and solution for finite element analysis, numerical linear algebra.", "model": "claude-haiku-5-5"}
++{"work_id": 3123748159, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Social-ecological knowledge pooling and fisheries modelling; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4406141980, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Pure mathematical analysis of nonlinear PDEs; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4310184906, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Parallelization and batched computation for RooFit statistical fitting software in high energy physics.", "model": "claude-haiku-5-5"}
++{"work_id": 3116267422, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Inverse problem analysis for photoacoustic tomography with numerical simulations; primarily applied mathematics.", "model": "claude-haiku-5-5"}
++{"work_id": 2810988304, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Monte Carlo regression algorithm for optimal stopping of McKean-Vlasov diffusions; mainly stochastic control and numerical methods, low-confidence fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4414647119, "categories": ["federated_learning", "distributed_computing"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Client sampling and grouping in multi-model federated learning with non-i.i.d. data is core FL work, with a distributed training systems aspect.", "model": "claude-haiku-5-5"}
++{"work_id": 2977976693, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Building-physics study of sky luminance distribution for natural illumination; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4286224978, "categories": ["federated_learning", "ml_theory"], "primary": "federated_learning", "confidence": 0.8, "rationale": "One-shot distributed learning via bootstrap refitting with statistical error rate analysis; federated/distributed statistical learning with theoretical focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4401880100, "categories": ["distributed_computing", "robotics"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Decentralised asynchronous ADMM optimisation algorithm for networked agents, applied to phase retrieval; primarily distributed optimisation.", "model": "claude-haiku-5-5"}
++{"work_id": 4403791687, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Contribution assessment of participants in federated learning without validation data.", "model": "claude-haiku-5-5"}
++{"work_id": 4411551389, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Adaptive local training epochs for federated learning under Non-IID data in IoV settings.", "model": "claude-haiku-5-5"}
++{"work_id": 4385602851, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Geophysical ionosphere electron density study of a solar eclipse; space science, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4404238664, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Weighted federated learning for heterogeneous clients in edge computing with non-IID data and model heterogeneity.", "model": "claude-haiku-5-5"}
++{"work_id": 2921727985, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Earth radiation observatory irradiance modelling from the Moon; remote sensing and atmospheric physics, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4390986554, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.75, "rationale": "Incentive mechanism design for fast federated learning with quality-aware participant selection; FL-centred, with economic mechanism aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 4404743912, "categories": ["federated_learning", "computer_vision", "computational_biology"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Federated learning for MRI image-to-image translation across sites; medical imaging is computer vision and biomedical application.", "model": "claude-haiku-5-5"}
++{"work_id": 4407950319, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Computing education study on challenge-based learning; not a CS research contribution in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 4415473342, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Federated training of a mask recognition model with client configuration tuning.", "model": "claude-haiku-5-5"}
++{"work_id": 4416017477, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Asynchronous personalized federated learning with communication sparsification.", "model": "claude-haiku-5-5"}
++{"work_id": 4313889858, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Federated learning framework with DenseNet for chest X-ray disease classification.", "model": "claude-haiku-5-5"}
++{"work_id": 7148322461, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Studies predictability and self-organization in deterministic cellular automata; closest to ML theory/learnability though interdisciplinary physics framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7154578995, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Survey of federated learning applications in healthcare informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 3215093560, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Convolutional autoencoder with texture fusion for PolSAR terrain classification from remote sensing imagery.", "model": "claude-haiku-5-5"}
++{"work_id": 7124262753, "categories": ["federated_learning", "ai_agents"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Agent-based holonic hierarchical distributed deep learning framework with aggregation for privacy-focused training.", "model": "claude-haiku-5-5"}
++{"work_id": 4408110066, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Survey of federated learning for Alzheimer's disease detection from medical data.", "model": "claude-haiku-5-5"}
++{"work_id": 7213405981, "categories": ["federated_learning", "ml_theory"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Core contribution is a federated optimization method addressing data heterogeneity.", "model": "claude-haiku-5-5"}
++{"work_id": 4413395786, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Federated training for ECG-based atrial fibrillation detection; medical application with FL as central method.", "model": "claude-haiku-5-5"}
++{"work_id": 4405697278, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Vertical federated deep learning for integrated IoT services.", "model": "claude-haiku-5-5"}
++{"work_id": 7131434230, "categories": ["federated_learning", "database_systems"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Overview of data analytics, big data and federated learning for business decision making.", "model": "claude-haiku-5-5"}
++{"work_id": 4400909753, "categories": ["federated_learning", "information_retrieval"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Federated framework for trajectory recovery; trajectory data processing is the application domain.", "model": "claude-haiku-5-5"}
++{"work_id": 7125126928, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Addresses catastrophic forgetting under temporal drift in federated learning via client-side replay.", "model": "claude-haiku-5-5"}
++{"work_id": 4401508498, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Plugin for concept drift in federated learning clients.", "model": "claude-haiku-5-5"}
++{"work_id": 4415004105, "categories": ["federated_learning", "nlp"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Evaluates BERT-based federated learning for Chinese text classification under non-IID data.", "model": "claude-haiku-5-5"}
++{"work_id": 4414119830, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.5, "rationale": "Federated learning framework for construction worker safety monitoring with digital twins and edge testbed.", "model": "claude-haiku-5-5"}
++{"work_id": 3210955338, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.4, "rationale": "Cyber-social computing paper proposing a cloud-edge federated learning architecture; largely conceptual.", "model": "claude-haiku-5-5"}
++{"work_id": 7163589765, "categories": ["hci", "ai_agents"], "primary": "hci", "confidence": 0.5, "rationale": "AI-supported collaborative learning instruction for students; educational interaction study. AI is a scaffold rather than an agent focus, so hci is primary; education is not a listed category.", "model": "claude-haiku-5-5"}
++{"work_id": 4403284007, "categories": ["federated_learning", "computer_networks"], "primary": "federated_learning", "confidence": 0.55, "rationale": "Title indicates federated learning under communication constraints; abstract is empty, so classification relies on title alone.", "model": "claude-haiku-5-5"}
++{"work_id": 3007895144, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Statistical estimation of discrete distributions from untrusted batches with sample-optimality guarantees; learning theory.", "model": "claude-haiku-5-5"}
++{"work_id": 7084124479, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Modified federated learning framework with convergence analysis versus FedAvg and FedProx.", "model": "claude-haiku-5-5"}
++{"work_id": 3210763747, "categories": ["federated_learning", "ai_safety"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Client-side defense against model poisoning in FL; robustness/security of FL systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7204715907, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Qualitative study of students' perceptions of AI use and critical thinking; user-focused human-AI interaction in education.", "model": "claude-haiku-5-5"}
++{"work_id": 2907334421, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Power electronics system with MPPT and DC bus control for lunar rover; hardware-level energy system, not clearly computing, so low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4401891542, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Federated adversarial training for fault diagnosis across clients; FL is central, application is industrial fault diagnosis from signals.", "model": "claude-haiku-5-5"}
++{"work_id": 4376149513, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Survey of heterogeneity-aware federated learning on constrained devices.", "model": "claude-haiku-5-5"}
++{"work_id": 2908566518, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Position column on a conceptual cosmology/history of AI as a field; not a research contribution in a specific CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7106710552, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Personalized federated test-time adaptation under class imbalance; federated learning is the core, with test-time adaptation of visual domain-shift models.", "model": "claude-haiku-5-5"}
++{"work_id": 7143299875, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Community service training on Google Drive for teachers; not CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4390204018, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Incentive and revenue-sharing mechanisms for cross-silo federated learning collaboration.", "model": "claude-haiku-5-5"}
++{"work_id": 4254855164, "categories": [], "primary": null, "confidence": 0.6, "rationale": "General introductory book text on AI history; no specific CS research category clearly applies.", "model": "claude-haiku-5-5"}
++{"work_id": 4313919668, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Federated Bayesian network learning across sites for brain connectivity of MDD patients; federated method applied to neuroimaging biomedical data.", "model": "claude-haiku-5-5"}
++{"work_id": 4412004712, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Review of federated learning in no-code AI for BFSI data security.", "model": "claude-haiku-5-5"}
++{"work_id": 4393928184, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Adaptive client selection using data entropy and model divergence for efficient FL.", "model": "claude-haiku-5-5"}
++{"work_id": 7085089513, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Federated graph learning with edge-client collaborative networks and subtopology-aware aggregation.", "model": "claude-haiku-5-5"}
++{"work_id": 4381744345, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Compressed, Byzantine-robust federated learning framework with hard-threshold compressors.", "model": "claude-haiku-5-5"}
++{"work_id": 4320855546, "categories": ["federated_learning", "computer_networks"], "primary": "federated_learning", "confidence": 0.75, "rationale": "Diffusion strategy for FL with non-IID data over 6G device-to-device communication.", "model": "claude-haiku-5-5"}
++{"work_id": 4403353453, "categories": ["ml_theory", "information_retrieval"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Zero-shot learning with knowledge graph completion and knowledge representation; learning paradigm rather than clear category; borderline between ml_theory and knowledge-graph/retrieval-type work.", "model": "claude-haiku-5-5"}
++{"work_id": 7128377297, "categories": ["foundation_models", "deep_learning_architectures"], "primary": "foundation_models", "confidence": 0.8, "rationale": "Time series foundation model with patch Transformer, scaling and ablations for zero-shot forecasting.", "model": "claude-haiku-5-5"}
++{"work_id": 4402260667, "categories": ["foundation_models", "computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Geospatial foundation model applied to canopy height estimation from satellite imagery; remote-sensing image analysis is primary.", "model": "claude-haiku-5-5"}
++{"work_id": 7169519287, "categories": ["ai_agents", "generative_ai", "reinforcement_learning"], "primary": "ai_agents", "confidence": 0.3, "rationale": "Introductory book chapter spanning many AI topics; no single clear contribution, so only weakly assigned.", "model": "claude-haiku-5-5"}
++{"work_id": 4402145393, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Discussion of practical AI in Earth sciences and geospatial data science; perspective paper with weak category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4410550951, "categories": ["computer_vision", "speech_audio", "foundation_models"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Survey of self-supervised learning for EEG signals; EEG is physiological time-series signal, closest to signal/audio processing, with foundation model aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 7171486454, "categories": ["deep_learning_architectures", "computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Conformer-based masked autoencoder self-supervised on EEG and HRV for clinical HIE classification; biomedical application with architecture contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7116294541, "categories": ["ml_theory", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Continual and transfer learning for building thermal dynamics models under concept drift; modelling of physical systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4416183290, "categories": ["computer_vision", "scientific_computing"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Automated glacial lake mapping from Sentinel-2 imagery using a segmentation foundation model; remote-sensing image analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 7133207779, "categories": ["foundation_models", "edge_computing", "computational_biology"], "primary": "foundation_models", "confidence": 0.7, "rationale": "Continual fine-tuning of a Transformer foundation model on wearable sensor data with PEFT, enabling edge disease detection.", "model": "claude-haiku-5-5"}
++{"work_id": 3189938493, "categories": ["deep_learning_architectures", "speech_audio"], "primary": "speech_audio", "confidence": 0.55, "rationale": "Transfer learning for radio signal modulation recognition; signal processing application of deep networks, not a foundation model.", "model": "claude-haiku-5-5"}
++{"work_id": 4412605435, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Online multisource transfer learning for fault diagnosis; transfer-learning method for a diagnostic application, weakly fits ml/DL methods.", "model": "claude-haiku-5-5"}
++{"work_id": 7172007151, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Meta-learning for wearable health models; primarily a training method applied to biosignal monitoring.", "model": "claude-haiku-5-5"}
++{"work_id": 4404035916, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.6, "rationale": "Cross-domain aspect sentiment triplet extraction using generative model with prompts; classical NLP task framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7171698092, "categories": ["computer_vision", "ai_safety"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Robustness metric for pathology image encoders against confounders; medical image analysis with shortcut-learning robustness concern.", "model": "claude-haiku-5-5"}
++{"work_id": 4417300144, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Adaptive trunk-branch architecture for time series under task conflict with theoretical analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 7130402983, "categories": ["llm", "ai_agents", "software_engineering"], "primary": "llm", "confidence": 0.9, "rationale": "GLM-5 foundation LLM with agentic RL and coding capabilities.", "model": "claude-haiku-5-5"}
++{"work_id": 4389821395, "categories": ["multimodal_ai", "reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Uses vision-language models as reward sources for RL agents; combines VLMs with RL.", "model": "claude-haiku-5-5"}
++{"work_id": 7203492078, "categories": ["multimodal_ai", "foundation_models"], "primary": "foundation_models", "confidence": 0.5, "rationale": "EEG foundation model with pretraining and instruction tuning, using text prototypes; neural signal decoding.", "model": "claude-haiku-5-5"}
++{"work_id": 4386106645, "categories": ["llm", "foundation_models", "distributed_computing"], "primary": "foundation_models", "confidence": 0.6, "rationale": "OpenBMB suite for training, tuning, compressing and serving big pretrained models.", "model": "claude-haiku-5-5"}
++{"work_id": 4407186881, "categories": ["foundation_models", "ml_theory"], "primary": "foundation_models", "confidence": 0.6, "rationale": "Improves TabPFN, a pre-trained in-context tabular foundation model, via fine-tuning and bagging; bias-variance framing is methodological.", "model": "claude-haiku-5-5"}
++{"work_id": 4387929345, "categories": ["nlp"], "primary": "nlp", "confidence": 0.75, "rationale": "Semi-supervised few-shot text classification of disaster tweets, a classical NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 4315588764, "categories": ["foundation_models"], "primary": "foundation_models", "confidence": 0.8, "rationale": "Self-supervised pretrained time-to-event foundation model for EHR data; non-language foundation model applied to medicine.", "model": "claude-haiku-5-5"}
++{"work_id": 7167906451, "categories": ["llm"], "primary": "llm", "confidence": 0.85, "rationale": "Continual parameter-efficient fine-tuning of 7-8B LLM backbones with LoRA-style adapters.", "model": "claude-haiku-5-5"}
++{"work_id": 7140202126, "categories": ["computational_biology", "computer_vision"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Benchmark deriving health metrics from PPG biosignals with deep learning; biomedical signal analysis, arguably health informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 4412420015, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Few-shot partial multi-label learning method with noisy label handling; general learning methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 4385757412, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Semi-supervised CNN autoencoder for ECG-based sleep apnea detection; biomedical health informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 4390871840, "categories": ["ai_safety", "foundation_models"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Studies bias inheritance from pretrained models in finetuning and its correction; fairness framed as safety problem.", "model": "claude-haiku-5-5"}
++{"work_id": 3160088594, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Few-shot domain adaptation for CT-based COVID-19 image diagnosis; medical image analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 7165377962, "categories": ["foundation_models"], "primary": "foundation_models", "confidence": 0.6, "rationale": "Zero-shot time series foundation models with covariates for building energy forecasting; sustainable application of foundation models.", "model": "claude-haiku-5-5"}
++{"work_id": 4399458826, "categories": ["computer_vision", "ml_theory"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Gait recognition from silhouettes/skeletons is computer vision; domain adaptation across viewing angles is the main contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7165063713, "categories": ["foundation_models", "ml_theory"], "primary": "foundation_models", "confidence": 0.4, "rationale": "Multi-source transfer learning and parameter merging across pretrained models for continual learning; not clearly foundation model specific.", "model": "claude-haiku-5-5"}
++{"work_id": 4313024900, "categories": ["deep_learning_architectures", "computer_architecture"], "primary": "deep_learning_architectures", "confidence": 0.3, "rationale": "Unsupervised domain adaptation for radar-based deep learning classification; application to radar sensing, weakly fits any category, closest to ML methods.", "model": "claude-haiku-5-5"}
++{"work_id": 7214033248, "categories": ["foundation_models", "computer_vision"], "primary": "foundation_models", "confidence": 0.5, "rationale": "Adapts a frozen pretrained MAE vision backbone to time-series forecasting, transfer of vision foundation model across domains.", "model": "claude-haiku-5-5"}
++{"work_id": 7138441586, "categories": ["llm", "foundation_models"], "primary": "llm", "confidence": 0.6, "rationale": "Parameter-efficient fine-tuning of large pretrained language models (LoRA/DoRA) evaluated on NLU and NLG.", "model": "claude-haiku-5-5"}
++{"work_id": 4402355991, "categories": ["computer_vision", "speech_audio"], "primary": "speech_audio", "confidence": 0.4, "rationale": "Representation learning on metro interior noise signals using VAE and clustering; audio signal analysis for fault detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4391766533, "categories": ["computer_vision", "foundation_models"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Self-supervised pretraining for whole slide histopathology images, visual representation learning in medical imaging.", "model": "claude-haiku-5-5"}
++{"work_id": 4405173494, "categories": ["ai_agents", "llm", "robotics"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Retrieval-augmented generalist agent acting in-context in robotics and game environments.", "model": "claude-haiku-5-5"}
++{"work_id": 7130707312, "categories": ["foundation_models", "computational_biology"], "primary": "foundation_models", "confidence": 0.5, "rationale": "EEG foundation model pretrained at scale with self-supervised objectives; biomedical signal modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 4403992746, "categories": ["foundation_models"], "primary": "foundation_models", "confidence": 0.3, "rationale": "Selecting pre-trained models for transfer learning using meta-features; title-only abstract, limited evidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4413751397, "categories": ["computational_biology", "deep_learning_architectures", "nlp", "speech_audio", "computer_vision", "multimodal_ai", "llm", "foundation_models"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Survey of non-intrusive deep learning for cognitive decline detection using speech, text, and visual modalities; health informatics fits computational_biology, with multimodal and foundation-model aspects. Trimmed to the most central categories.", "model": "claude-haiku-5-5"}
++{"work_id": 3134294021, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Event detection in English text data using transfer learning is an information extraction NLP task.", "model": "claude-haiku-5-5"}
++{"work_id": 3139784666, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Transfer learning and knowledge graph framework for sharing ML models across smart environments; closest fit is a learning-methodology framing, though it is more of a systems/IoT proposal.", "model": "claude-haiku-5-5"}
++{"work_id": 7167230086, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "CLIP-based multimodal zero-shot text classification with cross-lingual extension; vision-language plus multilingual NLP.", "model": "claude-haiku-5-5"}
++{"work_id": 4417127389, "categories": ["multimodal_ai", "deep_learning_architectures"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Pruning and low-rank adaptation of CLIP vision-language models for efficient fine-tuning and retrieval.", "model": "claude-haiku-5-5"}
++{"work_id": 7167035799, "categories": ["computer_vision", "computational_biology", "foundation_models"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Endoscopy foundation model for gastric cancer image analysis; medical image analysis is computer vision, with foundation model pretraining.", "model": "claude-haiku-5-5"}
++{"work_id": 6966932760, "categories": ["foundation_models", "generative_ai"], "primary": "foundation_models", "confidence": 0.5, "rationale": "Scaling laws and synthetic data generation for training large models; synthetic image generation for ImageNet training.", "model": "claude-haiku-5-5"}
++{"work_id": 7161153937, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "CNN-based breast cancer detection in mammograms; medical image analysis with preprocessing study.", "model": "claude-haiku-5-5"}
++{"work_id": 3126536330, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Domain adaptation network for bearing fault diagnosis from vibration/signal data; the applied focus is fault diagnosis, with transfer learning methods.", "model": "claude-haiku-5-5"}
++{"work_id": 7140216482, "categories": ["deep_learning_architectures", "llm", "multimodal_ai"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Systems and kernel optimization of DoRA low-rank adaptation for large models, including VLMs; fine-tuning method efficiency.", "model": "claude-haiku-5-5"}
++{"work_id": 4312223859, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Countermeasure against power side-channel attacks modeled via channel capacity; cryptographic hardware security framing, with security focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4391216411, "categories": ["sustainable_computing"], "primary": null, "confidence": 0.2, "rationale": "Power system capacity planning for low-carbon generation; not computer science research, so no category fits well.", "model": "claude-haiku-5-5"}
++{"work_id": 4391582496, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Image encryption scheme using chaotic systems, S-boxes and DNA coding; a cryptographic primitive design.", "model": "claude-haiku-5-5"}
++{"work_id": 4280508268, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Cryptanalysis of lightweight cipher TinyJAMBU with distinguishers and key-recovery attacks.", "model": "claude-haiku-5-5"}
++{"work_id": 4360859090, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Block-cipher-style image encryption algorithm combining chaos and DNA operations.", "model": "claude-haiku-5-5"}
++{"work_id": 3116994096, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Public-key encryption scheme over twisted Edwards curves with security proofs and signature variant.", "model": "claude-haiku-5-5"}
++{"work_id": 4400493258, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.75, "rationale": "Construction of cryptographic S-boxes with nonlinearity evaluation for symmetric encryption.", "model": "claude-haiku-5-5"}
++{"work_id": 7199797550, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.55, "rationale": "Bachelor thesis surveying cryptographic primitives for e-voting, with homomorphic encryption prototype.", "model": "claude-haiku-5-5"}
++{"work_id": 7016138925, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Thesis on side-channel evaluation and masking/shuffling countermeasure design for cryptographic implementations.", "model": "claude-haiku-5-5"}
++{"work_id": 3176687530, "categories": ["privacy_preserving_computation", "blockchain"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Privacy-preserving EHR data sharing on cloud with blockchain and IPFS; privacy is the main contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3135040192, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "GPU acceleration of NIST SP 800-90B entropy estimation for RNG noise sources; cryptographic module validation context, with parallel GPU implementation aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 4255230922, "categories": ["cybersecurity", "applied_cryptography"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Correlation power side-channel attack on smart card cryptographic devices.", "model": "claude-haiku-5-5"}
++{"work_id": 4312999490, "categories": ["cloud_computing", "cybersecurity", "quantum_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Survey of cloud computing security and the threat of quantum computing to it.", "model": "claude-haiku-5-5"}
++{"work_id": 7127423155, "categories": ["blockchain", "applied_cryptography"], "primary": "blockchain", "confidence": 0.7, "rationale": "Framework integrating self-signed key cryptography with blockchain, DIDs and IPFS for record authentication.", "model": "claude-haiku-5-5"}
++{"work_id": 4213389055, "categories": ["cybersecurity", "applied_cryptography"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Mutual assistant side-channel attack on AES using partial key recovery and genetic algorithm CPA.", "model": "claude-haiku-5-5"}
++{"work_id": 7108084985, "categories": ["quantum_computing", "applied_cryptography"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Book chapter on quantum random number generators for cryptographic randomness.", "model": "claude-haiku-5-5"}
++{"work_id": 4246586608, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Symmetric encryption algorithm using genetic crossover and mutation.", "model": "claude-haiku-5-5"}
++{"work_id": 4401369144, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Operations research and marketing model of repositioning under Markov regime switching; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4413953789, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.5, "rationale": "Practical implementation of post-quantum isogeny-based cryptography; abstract is only a publisher boilerplate, so judged from title.", "model": "claude-haiku-5-5"}
++{"work_id": 4401634049, "categories": ["applied_cryptography", "blockchain"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Threshold blind signature scheme based on Paillier, applied to blockchain transactions.", "model": "claude-haiku-5-5"}
++{"work_id": 3011920068, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Reliability of MLC NAND flash memory via constraint and error-correcting codes; storage hardware reliability, closest to computer architecture/memory systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2963804406, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.9, "rationale": "Enhancing security of the GGH lattice-based cryptosystem against lattice reduction attacks.", "model": "claude-haiku-5-5"}
++{"work_id": 4200321003, "categories": ["cybersecurity", "applied_cryptography"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Side-channel (CPA) attack on the SPECK lightweight cipher on Arduino; cryptanalysis via implementation attack.", "model": "claude-haiku-5-5"}
++{"work_id": 4401694671, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Securing healthcare cloud data with an LSTM-based optimized classifier; mixes cloud security with machine learning for image pixel classification.", "model": "claude-haiku-5-5"}
++{"work_id": 7215253795, "categories": ["blockchain", "applied_cryptography"], "primary": "blockchain", "confidence": 0.85, "rationale": "Blockchain-based decentralized platform with smart contracts and IPFS for digital credentials.", "model": "claude-haiku-5-5"}
++{"work_id": 2922278162, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Chat application encrypting database messages with Caesar Cipher and 3DES.", "model": "claude-haiku-5-5"}
++{"work_id": 4417415103, "categories": ["privacy_preserving_computation", "federated_learning", "blockchain"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Federated learning with differential privacy and Paillier homomorphic encryption for medical data sharing, plus blockchain trust.", "model": "claude-haiku-5-5"}
++{"work_id": 3157885010, "categories": ["applied_cryptography", "computer_architecture"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Hardware structure for NTT modular reduction used in post-quantum cryptography; area-efficient VLSI design.", "model": "claude-haiku-5-5"}
++{"work_id": 4251764918, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.75, "rationale": "Overview chapter of basic cryptographic methods, hash functions and public-key schemes.", "model": "claude-haiku-5-5"}
++{"work_id": 4416234071, "categories": ["applied_cryptography", "computational_complexity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Boolean function algebraic normal form monomial density bounds with cryptanalytic test applications to stream ciphers.", "model": "claude-haiku-5-5"}
++{"work_id": 2785483644, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Fuzzy vault biometric template protection, a bio-cryptographic primitive with chaff generation performance; security of biometric systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3126856290, "categories": ["computer_architecture", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Volt/VAR control in power distribution networks with optimisation and co-simulation; power systems engineering, not clearly a CS category, closest fit is scientific/engineering computing.", "model": "claude-haiku-5-5"}
++{"work_id": 3204643661, "categories": ["applied_cryptography", "privacy_preserving_computation"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Single-server private information retrieval scheme built on somewhat homomorphic encryption; cryptographic protocol design.", "model": "claude-haiku-5-5"}
++{"work_id": 4226498265, "categories": ["scientific_computing", "ml_theory"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Model reduction and MPC-based optimal irrigation scheduling for Richards equation PDE model; numerical/control methods for agro-hydrology.", "model": "claude-haiku-5-5"}
++{"work_id": 4405179471, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Historical biography of Enigma codebreakers; not CS research contribution in the ontology sense, though history of cryptography.", "model": "claude-haiku-5-5"}
++{"work_id": 7164908005, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Extended linear secure secret sharing for group key management; cryptographic primitive design and key management.", "model": "claude-haiku-5-5"}
++{"work_id": 4404987385, "categories": ["applied_cryptography", "quantum_computing"], "primary": "quantum_computing", "confidence": 0.75, "rationale": "Source-independent continuous-variable quantum random number generator with security proof; quantum information processing for cryptographic randomness.", "model": "claude-haiku-5-5"}
++{"work_id": 4391590938, "categories": ["applied_cryptography", "deep_learning_architectures", "distributed_computing"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "GPU-accelerated secure multi-party computation framework for neural network training and Transformer inference; MPC primitives with ML application.", "model": "claude-haiku-5-5"}
++{"work_id": 4392426149, "categories": ["applied_cryptography", "privacy_preserving_computation"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Two-party secure computation protocols for nonlinear activation functions in neural networks; cryptographic protocol design.", "model": "claude-haiku-5-5"}
++{"work_id": 4391096928, "categories": ["applied_cryptography", "blockchain"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Survey of post-quantum exotic signature schemes motivated by blockchain security; primarily cryptographic primitives.", "model": "claude-haiku-5-5"}
++{"work_id": 4401247688, "categories": ["cloud_computing", "applied_cryptography", "blockchain"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Integrity audit/proof of retrievability for cloud and edge stored data with blockchain context; primarily a cryptographic/integrity verification scheme, so primary is cybersecurity.", "model": "claude-haiku-5-5"}
++{"work_id": 4200232553, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Commentary on transition to post-quantum cryptographic standards.", "model": "claude-haiku-5-5"}
++{"work_id": 7160766211, "categories": ["applied_cryptography", "distributed_computing"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Theoretical framework for secure multi-party computation with dynamic committees.", "model": "claude-haiku-5-5"}
++{"work_id": 7220773489, "categories": ["applied_cryptography", "computer_architecture"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Performance evaluation of the SPECK block cipher on Raspberry Pi hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4402836518, "categories": ["applied_cryptography", "computer_vision"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Chaotic hyperchaotic image encryption scheme using DNA coding.", "model": "claude-haiku-5-5"}
++{"work_id": 4378805032, "categories": ["computer_architecture"], "primary": null, "confidence": 0.3, "rationale": "Power systems control of EV charging stations in microgrids; not clearly a CS category, though closest is embedded/control hardware. Conservatively low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 3073667118, "categories": ["applied_cryptography", "computational_complexity"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Key exchange protocol construction over non-commutative groups with security tied to an NP-complete problem.", "model": "claude-haiku-5-5"}
++{"work_id": 3090466325, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Constructing finite field elements of provably high multiplicative order; algebraic/computational number theory with cryptographic motivation.", "model": "claude-haiku-5-5"}
++{"work_id": 4402426511, "categories": ["quantum_computing", "applied_cryptography"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Thesis on bosonic quantum key distribution and quantum error-correcting codes.", "model": "claude-haiku-5-5"}
++{"work_id": 4392502209, "categories": ["applied_cryptography", "ai_accelerators"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "GPU-parallel implementation of the SPHINCS+ post-quantum signature scheme.", "model": "claude-haiku-5-5"}
++{"work_id": 3115113005, "categories": ["computational_complexity", "computer_graphics"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Introduces a computational model (fragment shader machines) and studies computable functions and running times; uses GPU fragment shaders, so graphics hardware is involved.", "model": "claude-haiku-5-5"}
++{"work_id": 4232616731, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Book on data visualization and information design for communicating data to audiences; design-oriented and not clearly CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4381744354, "categories": ["information_retrieval", "hci"], "primary": "hci", "confidence": 0.3, "rationale": "Business intelligence dashboard design with computer-supported cooperative work framing; primarily visualization and CSCW design for an industry platform.", "model": "claude-haiku-5-5"}
++{"work_id": 3185991883, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Image-based rendering and view synthesis for non-Lambertian light fields.", "model": "claude-haiku-5-5"}
++{"work_id": 2909271333, "categories": ["computer_graphics", "algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Formalization of spatial configurations and packing, layout and covering constraints, a combinatorial and geometric optimization problem.", "model": "claude-haiku-5-5"}
++{"work_id": 4413330002, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.45, "rationale": "Workshop on esports rendering needs with developers and players, combining rendering and user perspectives.", "model": "claude-haiku-5-5"}
++{"work_id": 4382541407, "categories": ["distributed_computing", "scientific_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Hybrid MPI+OpenMP/OpenACC parallelization and dynamic load balancing of a DSMC scientific solver on HPC systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4285816232, "categories": ["computer_graphics", "distributed_computing", "cloud_computing"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "Parallel processing acceleration of complex 3D scene rendering in a B/S cloud setting.", "model": "claude-haiku-5-5"}
++{"work_id": 4400878583, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "3D virtual modeling and color rendering from hand-drawn strokes for graphic design.", "model": "claude-haiku-5-5"}
++{"work_id": 3035417602, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Indonesian public-service advertisement audiovisual design; communication design, not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7139845931, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Restoration of distance matrices and algorithms for reconstructing Euclidean distance matrices; mathematical/algorithmic focus, with applications mentioned but secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 3044538714, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.85, "rationale": "Neural scene representation for free-viewpoint rendering with sparse voxel octrees; neural rendering is a computer graphics topic.", "model": "claude-haiku-5-5"}
++{"work_id": 4249339162, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Mesh improvement and smoothing for geometry processing and modelling, a graphics/geometry topic; could also be scientific computing.", "model": "claude-haiku-5-5"}
++{"work_id": 2758763654, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Augmented reality application for drum instrument visualization with evaluation of camera parameters; interactive system study.", "model": "claude-haiku-5-5"}
++{"work_id": 2785914471, "categories": ["hci", "computer_vision"], "primary": "hci", "confidence": 0.5, "rationale": "Assessing clutter in parallel coordinates visualization using image processing; visualization and user-facing evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 4414732984, "categories": ["generative_ai", "multimodal_ai", "speech_audio"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Generates audio-visual art from colors and chords with an LSTM; cross-modal generation involving audio and images.", "model": "claude-haiku-5-5"}
++{"work_id": 4389429756, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Art project on AI-generated portraits and cosmology; not primarily computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7151078784, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Evaluation of easing functions for game animation in a game engine; graphics/animation topic.", "model": "claude-haiku-5-5"}
++{"work_id": 4229820775, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Popular history book about DOS games and DirectX; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4417339164, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.7, "rationale": "Neural implicit distance field reconstruction of dynamic urban scenes and objects; 3D reconstruction and neural representations.", "model": "claude-haiku-5-5"}
++{"work_id": 4409753672, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Visual identity design for an art market in Indonesian; design research, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7163058269, "categories": ["hci", "information_retrieval"], "primary": "hci", "confidence": 0.6, "rationale": "Visual analytics strategies for coordinating DFG and timeline views is interactive visualization for process analysts; primarily HCI/visual analytics, process mining related.", "model": "claude-haiku-5-5"}
++{"work_id": 4382563412, "categories": ["hci", "computer_graphics"], "primary": "hci", "confidence": 0.6, "rationale": "Prototype representing amphibian sensory perception in a Unity first-person game; interactive/experiential design with game-engine implementation.", "model": "claude-haiku-5-5"}
++{"work_id": 2930742311, "categories": ["computer_graphics", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Numerical SPH multiphase fluid simulation of landslide waves with visualization; computational physics simulation.", "model": "claude-haiku-5-5"}
++{"work_id": 2340447824, "categories": ["computer_architecture", "computer_graphics"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Hardware architecture for real-time volume rendering with parallel pipeline stages.", "model": "claude-haiku-5-5"}
++{"work_id": 4404698773, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Evaluates AI video generation platforms (Runway, Luma) for producing moving images; communication studies framing but generative video focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3183006928, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.7, "rationale": "Scientific visualization and physically-based rendering application built on OSPRay.", "model": "claude-haiku-5-5"}
++{"work_id": 7167851020, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "Production session describing animation/rendering pipeline for a feature film; graphics production, not strictly research.", "model": "claude-haiku-5-5"}
++{"work_id": 7206470547, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Abstract only says 66 pages; data modelling textbook in Spanish, insufficient info to assign computer science category confidently.", "model": "claude-haiku-5-5"}
++{"work_id": 4400958279, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Real-time stereo rendering technique improving screen space reflections.", "model": "claude-haiku-5-5"}
++{"work_id": 4404548282, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Real-time rendering of 3D virtual scenes with raindrop simulation on a virtual lens, using GPU parallelism.", "model": "claude-haiku-5-5"}
++{"work_id": 4412588572, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.75, "rationale": "Optimization of procedural patterns with user editing, a graphics modelling contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 2966003478, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Study of polygon reduction effects on 3D mesh models for CG artists.", "model": "claude-haiku-5-5"}
++{"work_id": 7218025575, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.55, "rationale": "Comparison of procedural maze generation algorithms for game content; closest fit is procedural content in graphics, though algorithmic in nature.", "model": "claude-haiku-5-5"}
++{"work_id": 4375835151, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Production of a 3D promotional video for a property development; marketing media work, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4280553375, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.7, "rationale": "Quad-fisheye image stitching for panorama reconstruction via mesh deformation, with real-time CUDA implementation.", "model": "claude-haiku-5-5"}
++{"work_id": 3037385141, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "Visualization technique for spatial multi-hazard risk data using convex hull regions and color schemes.", "model": "claude-haiku-5-5"}
++{"work_id": 2789701916, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Streamgraph visualization of hierarchical time series with focus+context exploration.", "model": "claude-haiku-5-5"}
++{"work_id": 4392521761, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Development of a 3D animation video for public education about ventilators; multimedia production rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7221345862, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.75, "rationale": "Ray-tracing-based image distortion and blending correction for curved VR projection screens.", "model": "claude-haiku-5-5"}
++{"work_id": 4296604574, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Reflections on running creative visualization learning activities in visualization education; educational/HCI-adjacent, not clearly a specific category; assigned hci as closest fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4393211575, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.4, "rationale": "French paper on simplicial and cellular structures for geometric modelling; abstract is minimal, geometric modelling suggests computer_graphics.", "model": "claude-haiku-5-5"}
++{"work_id": 4385490078, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Non-photorealistic rendering system for conveying shape with suggestive contours; rendering is computer_graphics.", "model": "claude-haiku-5-5"}
++{"work_id": 4321788308, "categories": ["information_retrieval", "database_systems"], "primary": "database_systems", "confidence": 0.35, "rationale": "Visualization system for supermarket transaction data mining to show related goods; data analytics on purchase records.", "model": "claude-haiku-5-5"}
++{"work_id": 2911494573, "categories": ["distributed_computing", "computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Sparse matrix-vector multiplication auto-tuner for GPGPUs; hardware-aware performance optimisation, parallel HPC software.", "model": "claude-haiku-5-5"}
++{"work_id": 4405709446, "categories": ["generative_ai", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.5, "rationale": "AIGC integration for digital humans in the metaverse, multimodal interaction and content generation.", "model": "claude-haiku-5-5"}
++{"work_id": 4220934601, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.3, "rationale": "VR and 3D rendering for print advertisement visual optimisation; loosely graphics with image mining.", "model": "claude-haiku-5-5"}
++{"work_id": 4417539761, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.7, "rationale": "Articulated 3D object modelling with 3D Gaussian splatting for reconstruction and motion; neural 3D representations.", "model": "claude-haiku-5-5"}
++{"work_id": 2804059852, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Phase error correction for fringe projection 3D profilometry, a 3D reconstruction vision method.", "model": "claude-haiku-5-5"}
++{"work_id": 4417124644, "categories": ["generative_ai", "computer_graphics"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Generative 4D inbetweening from two images using video interpolation and multi-view diffusion.", "model": "claude-haiku-5-5"}
++{"work_id": 7131893792, "categories": ["multimodal_ai", "robotics"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Uses LLaVA vision-language model with prompt engineering for pedestrian behavior detection from robot perspective; primarily VLM work with robotics navigation context.", "model": "claude-haiku-5-5"}
++{"work_id": 7160568041, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Benchmark for vision-language chart understanding and hallucination detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4417286526, "categories": ["multimodal_ai", "hci"], "primary": "multimodal_ai", "confidence": 0.55, "rationale": "Evaluates VLMs for driver monitoring systems; application of vision-language models to in-cabin monitoring.", "model": "claude-haiku-5-5"}
++{"work_id": 3157008771, "categories": ["ml_theory", "information_retrieval"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Rank correlation coefficients, decision trees and ensembles for ranking data, plus linear mixed model review; statistical methodology.", "model": "claude-haiku-5-5"}
++{"work_id": 3212108239, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Continual learning of language grounding in CLIP embeddings for referring expression tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 7204869979, "categories": ["multimodal_ai", "ai_agents"], "primary": "multimodal_ai", "confidence": 0.55, "rationale": "Manga visual question answering with active panel selection and RL-style training over multimodal narratives.", "model": "claude-haiku-5-5"}
++{"work_id": 4402703549, "categories": ["multimodal_ai", "foundation_models"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Open large multimodal models family with training recipe and datasets; vision-language model release.", "model": "claude-haiku-5-5"}
++{"work_id": 7083693837, "categories": ["multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Benchmark evaluating vision-language models on multimodal clinical reasoning in neurology.", "model": "claude-haiku-5-5"}
++{"work_id": 4324302630, "categories": ["multimodal_ai", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.5, "rationale": "Applies CLIP multimodal models to explore digital visual historical collections in digital humanities.", "model": "claude-haiku-5-5"}
++{"work_id": 7172396779, "categories": ["multimodal_ai", "computer_vision"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Training-free text-image state localization in object video histories using frozen vision-language encoder.", "model": "claude-haiku-5-5"}
++{"work_id": 3176349206, "categories": ["multimodal_ai", "generative_ai", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Cross-modal retrieval with GANs and graph networks over image-text data; multimodal retrieval is the core task.", "model": "claude-haiku-5-5"}
++{"work_id": 7220827782, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Abstract is only a venue placeholder; the title suggests RAG/multimodal retrieval for nutrition, but insufficient content to be confident.", "model": "claude-haiku-5-5"}
++{"work_id": 4412945255, "categories": ["multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Evaluates multimodal LLMs on medical imaging with compositional generalization over vision-language tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 7197818024, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.6, "rationale": "PhD dissertation on multimodal deep learning for medical data, classifying diabetic retinopathy from images.", "model": "claude-haiku-5-5"}
++{"work_id": 4415538349, "categories": ["multimodal_ai", "llm", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Multimodal LLM with dialogue graphs for multimodal conversational aspect-based sentiment analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4403884255, "categories": ["multimodal_ai", "llm", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Open-source large multimodal model acting as evaluator and reward source for preference learning and alignment.", "model": "claude-haiku-5-5"}
++{"work_id": 7134939474, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Multimodal sentiment analysis fusing text, audio and visual cues with cross-modal attention.", "model": "claude-haiku-5-5"}
++{"work_id": 3205782068, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Audio-visual scene-aware dialog with AV-transformers and temporal reasoning over video.", "model": "claude-haiku-5-5"}
++{"work_id": 7166830964, "categories": ["multimodal_ai", "llm", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Evaluates large vision-language models on Bangla medical visual question answering.", "model": "claude-haiku-5-5"}
++{"work_id": 4394564345, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Visual dialog with cross-modal interaction and graph reasoning over image and dialogue history.", "model": "claude-haiku-5-5"}
++{"work_id": 4381956051, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal sentiment analysis using vision-language pre-trained model; the safety tag is not warranted, so only multimodal_ai should be kept as the main category.", "model": "claude-haiku-5-5"}
++{"work_id": 7214149876, "categories": ["multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Benchmarks and uses MLLMs for chart image reconstruction, a vision-language task with an interactive system.", "model": "claude-haiku-5-5"}
++{"work_id": 7160273623, "categories": ["multimodal_ai", "speech_audio"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal corpus of audio, video, motion and language for action verb learning in human-robot interaction; a dataset contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4416060734, "categories": ["multimodal_ai", "generative_ai"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Diffusion-based generation of missing modality features for multimodal emotion recognition across speech, text and vision.", "model": "claude-haiku-5-5"}
++{"work_id": 7106312846, "categories": ["multimodal_ai", "computer_vision", "nlp"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Chest X-ray report generation from images with CLIP-based encoders and text decoders, a vision-to-language task.", "model": "claude-haiku-5-5"}
++{"work_id": 4414169798, "categories": ["multimodal_ai", "computer_vision", "computational_biology"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Medical image captioning of chest X-rays with CheXNet and BioGPT; a vision-language task in a clinical setting.", "model": "claude-haiku-5-5"}
++{"work_id": 7117449861, "categories": ["multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Multi-hop visual chain-of-thought reasoning over frozen vision-language foundation models for VQA on litter images.", "model": "claude-haiku-5-5"}
++{"work_id": 4387968461, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Image captioning with knowledge injection from vision-language pre-trained models.", "model": "claude-haiku-5-5"}
++{"work_id": 7131242786, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Evaluation and comparison of open-source text-to-image diffusion models for artistic image generation.", "model": "claude-haiku-5-5"}
++{"work_id": 7155044497, "categories": ["speech_audio", "llm"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Speech summarization with multimodal LLMs trained via reinforcement learning; speech-to-text summarization of audio.", "model": "claude-haiku-5-5"}
++{"work_id": 4410698082, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal sentiment analysis with cross-modal attention and contrastive alignment; primary is multimodal_ai. Sentiment classification also touches nlp, but the core is multimodal fusion.", "model": "claude-haiku-5-5"}
++{"work_id": 7128505395, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Fake news video explanation generation aligning language and video features; multimodal video-text work with some NLP generation framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4387969592, "categories": ["generative_ai", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Image reconstruction from fMRI using Stable Diffusion; synthesis is generative and links brain signals with image modality.", "model": "claude-haiku-5-5"}
++{"work_id": 2998378979, "categories": ["multimodal_ai", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Cross-modal retrieval across multiple modalities with graph attention and adversarial learning.", "model": "claude-haiku-5-5"}
++{"work_id": 4404688739, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.5, "rationale": "Interpretability of image-text pairing for multimodal data fusion; multimodal focus with an interpretability angle.", "model": "claude-haiku-5-5"}
++{"work_id": 3201123210, "categories": ["multimodal_ai", "hci"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal emotion recognition fusing text, speech and visual inputs via graph convolution.", "model": "claude-haiku-5-5"}
++{"work_id": 4412082472, "categories": ["multimodal_ai", "deep_learning_architectures"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Token pruning in multimodal models like BLIP-2 for efficient inference; model compression on a vision-language model.", "model": "claude-haiku-5-5"}
++{"work_id": 4412438803, "categories": ["multimodal_ai", "generative_ai"], "primary": "multimodal_ai", "confidence": 0.5, "rationale": "Poster layout generation using text and image understanding with chain-of-thought; multimodal and generative layout synthesis.", "model": "claude-haiku-5-5"}
++{"work_id": 3034253961, "categories": ["ai_agents", "reinforcement_learning", "multimodal_ai"], "primary": "ai_agents", "confidence": 0.7, "rationale": "Vision-and-language navigation agent learned with imitation and curriculum RL; embodied agent in simulated environments.", "model": "claude-haiku-5-5"}
++{"work_id": 7213885858, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Robust multimodal image-text crisis classification under degraded modalities; multimodal classification with robustness focus.", "model": "claude-haiku-5-5"}
++{"work_id": 7126829820, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal meme classification with image and text embeddings; sexism detection relates to AI safety/bias framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7155094944, "categories": ["computer_vision", "ml_theory"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Abstract visual reasoning on RAVEN images with contrastive rule learning; primarily a neural reasoning architecture for visual puzzles.", "model": "claude-haiku-5-5"}
++{"work_id": 7169863823, "categories": ["multimodal_ai", "llm", "ai_agents"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Multimodal LLM adapted to vision-language navigation with visual encoder alignment; embodied navigation agent.", "model": "claude-haiku-5-5"}
++{"work_id": 4313013123, "categories": ["multimodal_ai", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Survey of multimodal event detection from Twitter; social media mining.", "model": "claude-haiku-5-5"}
++{"work_id": 4404450837, "categories": ["multimodal_ai", "generative_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Any-to-many multimodal LLM generating combinations of modalities.", "model": "claude-haiku-5-5"}
++{"work_id": 4283511515, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Image captioning and text features for multimodal cheapfake/misinformation detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4398157060, "categories": ["ai_agents", "llm", "multimodal_ai"], "primary": "llm", "confidence": 0.7, "rationale": "LLM-based vision-and-language navigation agent with memory map and reasoning chains.", "model": "claude-haiku-5-5"}
++{"work_id": 4405779000, "categories": ["computer_vision", "multimodal_ai"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Vision anomaly detection guided by language modality; cross-modal guidance for images.", "model": "claude-haiku-5-5"}
++{"work_id": 7114776455, "categories": ["generative_ai", "nlp", "hci"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Template-based text-to-image generation for cognitive accessibility from simplified text.", "model": "claude-haiku-5-5"}
++{"work_id": 7171869391, "categories": ["llm", "multimodal_ai", "computer_vision"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Augmenting multimodal LLMs with knowledge retrieval for substation image-based safety inspection.", "model": "claude-haiku-5-5"}
++{"work_id": 7118751186, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Library science survey of digital library services; not a clear CS research category.", "model": "claude-haiku-5-5"}
++{"work_id": 4321485607, "categories": ["information_retrieval", "deep_learning_architectures"], "primary": "information_retrieval", "confidence": 0.85, "rationale": "Two-stage code search and retrieval ranking with bi-encoders and cross-encoders.", "model": "claude-haiku-5-5"}
++{"work_id": 7124998404, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Recommendation system for streams and colleges; recommender systems framing with NLP interface.", "model": "claude-haiku-5-5"}
++{"work_id": 4376653350, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Scoping review of reviewer assignment as matching/recommendation problem.", "model": "claude-haiku-5-5"}
++{"work_id": 4285188099, "categories": ["ml_theory", "database_systems"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Ontology-based meta-learning recommendation of feature selection algorithms; semantic web and meta-learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7219176552, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.25, "rationale": "Only a thesis title-like abstract on skill recommendation; weak signal.", "model": "claude-haiku-5-5"}
++{"work_id": 7221366546, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Sentiment analysis of Indonesian comments with Naive Bayes.", "model": "claude-haiku-5-5"}
++{"work_id": 7135293231, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Business and information systems thesis on paid Q&A user engagement; not clearly a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7204876499, "categories": ["llm", "digital_libraries_placeholder_none"], "primary": "llm", "confidence": 0.5, "rationale": "Curatability of open LLM artifacts on Hugging Face and scholarly records; LLM-centred scholarly metadata study.", "model": "claude-haiku-5-5"}
++{"work_id": 3175543912, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Information science terminology study on exploratory search in Polish; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4319781162, "categories": ["information_retrieval", "llm"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Improves collaborative topic regression recommender for broad-perspective users; uses a pre-trained language model to augment vocabulary for document-based recommendation.", "model": "claude-haiku-5-5"}
++{"work_id": 3208454977, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Recommends datasets for research problem descriptions, a retrieval/recommendation task framed as text classification.", "model": "claude-haiku-5-5"}
++{"work_id": 4391682048, "categories": ["information_retrieval", "cybersecurity"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "Comparative study of privacy-aware recommender systems using perturbed ratings; recommendation is primary, privacy is the lens.", "model": "claude-haiku-5-5"}
++{"work_id": 7125905211, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "POI recommendation via CTR prediction with graph and Mamba encoders.", "model": "claude-haiku-5-5"}
++{"work_id": 2888686055, "categories": ["information_retrieval", "edge_computing"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Weather and location aware recommender using IoT sensors in a pervasive environment; recommendation is the core task.", "model": "claude-haiku-5-5"}
++{"work_id": 2798885959, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "User study of web search readability for dyslexic users; accessibility and human factors focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3093888250, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Only a funding statement is available; sequential user behavior modeling suggests recommendation, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7206221668, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library information bulletin, not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2907543384, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Ontology-based personalized learning path recommendation for e-learning.", "model": "claude-haiku-5-5"}
++{"work_id": 7138937744, "categories": ["information_retrieval", "multimodal_ai", "reinforcement_learning"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "E-commerce recommendation with multimodal fusion of text and images plus RL-based marketing decisions.", "model": "claude-haiku-5-5"}
++{"work_id": 3036207666, "categories": ["ml_theory", "information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Fairness framework for two-sided marketplaces such as search and recommender systems; fairness optimisation is a formal ML problem, with recommendation/search as the application setting.", "model": "claude-haiku-5-5"}
++{"work_id": 4409095591, "categories": ["information_retrieval", "computer_graphics"], "primary": "information_retrieval", "confidence": 0.45, "rationale": "Knowledge-graph-based classification of medical VR resources with a recommendation system; VR is the object domain but the contribution is classification and recommendation.", "model": "claude-haiku-5-5"}
++{"work_id": 4399261954, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.35, "rationale": "Web scraping of e-commerce product information for analysis; closest fit is web mining in information retrieval, though the abstract is very thin.", "model": "claude-haiku-5-5"}
++{"work_id": 4224219747, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Book chapter on latent factors in collaborative filtering, a recommender-system technique; abstract is largely marketing discussion.", "model": "claude-haiku-5-5"}
++{"work_id": 7172068451, "categories": ["information_retrieval", "software_engineering"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Neural code-to-code semantic retrieval system with FAISS index and benchmark against BM25; retrieval is the core contribution, applied to software repositories.", "model": "claude-haiku-5-5"}
++{"work_id": 7221540348, "categories": ["hci", "cybersecurity"], "primary": "hci", "confidence": 0.5, "rationale": "Child recognition from user image-selection patterns in web interaction; user behaviour study with a security/access-control motivation.", "model": "claude-haiku-5-5"}
++{"work_id": 7160413516, "categories": ["information_retrieval", "multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Vision-language models reasoning over document screenshots for visual attribution in iterative RAG; multimodal is the core, with retrieval-augmented generation context.", "model": "claude-haiku-5-5"}
++{"work_id": 4415775831, "categories": ["llm", "software_engineering"], "primary": "llm", "confidence": 0.6, "rationale": "RAG system using Mistral-7B with QLoRA to generate automated feedback on maritime assessments; LLM application in education.", "model": "claude-haiku-5-5"}
++{"work_id": 4306317060, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.85, "rationale": "Consistency metric for evaluating ranking models across query reformulations; IR evaluation of neural rankers.", "model": "claude-haiku-5-5"}
++{"work_id": 4409365993, "categories": ["ml_theory", "information_retrieval"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Multi-task learning optimisation via Lagrangian multiplier methods with theoretical guarantees, applied to web search ranking at Taobao.", "model": "claude-haiku-5-5"}
++{"work_id": 2883625840, "categories": ["information_retrieval", "ml_theory"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "Destination recommendation system using ML classifiers; recommender systems fall under information_retrieval, with decision-tree/ensemble modelling secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4224311926, "categories": ["information_retrieval", "ml_theory"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "Interpretable collaborative filtering for recommendation; recommender systems are primary, model-level interpretability as a method is secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 7131644180, "categories": ["information_retrieval", "llm"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "LLM-based offline reasoning for recommender reranking and item indexing; recommendation is primary with LLM use central.", "model": "claude-haiku-5-5"}
++{"work_id": 4404884507, "categories": ["privacy_preserving_computation", "information_retrieval"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Combines differential privacy and blockchain for privacy-preserving recommendation; privacy is the main focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3208034694, "categories": ["hci", "cybersecurity"], "primary": "hci", "confidence": 0.6, "rationale": "Automated accessibility evaluation simulating users with and without disabilities; accessibility testing of web interfaces.", "model": "claude-haiku-5-5"}
++{"work_id": 4377079764, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Multi-interest sequential recommendation model for consumer behavior.", "model": "claude-haiku-5-5"}
++{"work_id": 4412377349, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Graph debiasing for next POI recommendation, a recommender systems task.", "model": "claude-haiku-5-5"}
++{"work_id": 3094673112, "categories": ["information_retrieval", "ai_agents"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Survey of AI systems for tourism including recommendation and personal agents; broad applied survey with weak category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 7206653269, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library information bulletin, not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2801495198, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Book review of an internet searching guide for librarians, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3000081209, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.85, "rationale": "Film recommender system based on collaborative filtering is a recommender systems contribution, which the ontology places under information_retrieval.", "model": "claude-haiku-5-5"}
++{"work_id": 7128791697, "categories": ["information_retrieval", "multimodal_ai", "hci"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "Emotion-aware YouTube recommender using facial emotion detection; recommendation is primary, with affective computing and facial recognition secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4389683448, "categories": ["hci", "information_retrieval"], "primary": "hci", "confidence": 0.6, "rationale": "Survey study of consumer psychology and purchase intention regarding personalized recommendation systems; user-centred behavioural study.", "model": "claude-haiku-5-5"}
++{"work_id": 6910675515, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Bibliometric and sociology-of-science study of AI research norms; not a computer-science technical contribution in the ontology sense, so no category.", "model": "claude-haiku-5-5"}
++{"work_id": 4391150058, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Semantic web personalization model for web pages and search engines; loosely fits information_retrieval, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7220856794, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Practical theology literature review on recommendation algorithms and self-narrative; not computer-science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7217759194, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Category-based search with session clustering and user-interaction-based ranking; search personalization fits information_retrieval.", "model": "claude-haiku-5-5"}
++{"work_id": 7160261118, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Neural causal inference for individual treatment effect estimation; causal inference method fits ml_theory, though it could be seen as applied ML.", "model": "claude-haiku-5-5"}
++{"work_id": 4415702047, "categories": ["privacy_preserving_computation", "information_retrieval"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "Privacy-preserving representation learning defending attribute inference attacks in recommenders; privacy is primary, recommendation secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 7160178876, "categories": ["information_retrieval", "software_engineering"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Website hyperlink structure analysis and reconstruction using page similarity; web mining with some software/web engineering framing, low-moderate confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 3169195854, "categories": ["compilers"], "primary": "compilers", "confidence": 0.85, "rationale": "Loop transformations, polyhedral model and compiler optimization evaluation in CMSSW code; compiler optimization focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4412395283, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Clinical randomized trial of a mobile health app for blood pressure; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2798370950, "categories": ["compilers"], "primary": "compilers", "confidence": 0.9, "rationale": "Region-based profile-guided JIT compiler design for PHP/Hack; JIT compilation fits compilers.", "model": "claude-haiku-5-5"}
++{"work_id": 3179712771, "categories": [], "primary": null, "confidence": 0.8, "rationale": "ESL grammar learning resource from language education; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7213884990, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Community service training on writing literature reviews for high school students; not CS.", "model": "claude-haiku-5-5"}
++{"work_id": 1555598009, "categories": ["compilers"], "primary": "compilers", "confidence": 0.85, "rationale": "Code generator with intra-block stack scheduling eliminating redundant variable accesses; compiler code generation.", "model": "claude-haiku-5-5"}
++{"work_id": 7221269676, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "AutoLISP automation program for electrical SLD design; mainly an engineering tool/automation application, weakly software-related.", "model": "claude-haiku-5-5"}
++{"work_id": 2591985246, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.35, "rationale": "Computing education course reform integrating C into computer organization; mostly education, weak fit to programming languages.", "model": "claude-haiku-5-5"}
++{"work_id": 3109527096, "categories": ["compilers"], "primary": "compilers", "confidence": 0.85, "rationale": "Compiler optimization eliding redundant semiring expressions, evaluated on LLVM test suite.", "model": "claude-haiku-5-5"}
++{"work_id": 4407848518, "categories": ["compilers"], "primary": "compilers", "confidence": 0.9, "rationale": "MLIR Transform dialect for fine-grained control of compiler transformations and optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 4282571853, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "DC motor speed control with PID and fuzzy logic is control engineering, not clearly any listed CS category; weakly fits none well, so low-confidence placeholder with no categories.", "model": "claude-haiku-5-5"}
++{"work_id": 4410182237, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Educational puzzle game for early childhood using forward chaining; interactive system design and usability testing fit hci, with rule-based reasoning as the method.", "model": "claude-haiku-5-5"}
++{"work_id": 4321500346, "categories": ["compilers", "programming_languages"], "primary": "compilers", "confidence": 0.75, "rationale": "CARMOT is a PSEC-specific compiler co-designed with a parallel runtime for OpenMP and C++ abstraction recommendations.", "model": "claude-haiku-5-5"}
++{"work_id": 2908836625, "categories": ["database_systems"], "primary": null, "confidence": 0.3, "rationale": "Compilation of probabilistic business rules using PRM; abstract is empty, so only a weak guess toward rule/knowledge-base systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7128400621, "categories": ["compilers", "formal_verification"], "primary": "compilers", "confidence": 0.6, "rationale": "Transfers classical compiler optimization passes (constant propagation, dead-code elimination) to quantum programs via static analyses.", "model": "claude-haiku-5-5"}
++{"work_id": 4386528956, "categories": ["compilers", "distributed_computing"], "primary": "compilers", "confidence": 0.7, "rationale": "Direct GPU compilation of host applications via LLVM/OpenMP with ensemble execution to improve parallelism and GPU utilization.", "model": "claude-haiku-5-5"}
++{"work_id": 2901311712, "categories": ["compilers", "distributed_computing"], "primary": "compilers", "confidence": 0.55, "rationale": "GPU optimization via block size tuning and mixed precision tuning for CUDA kernels; performance tuning of parallel programs.", "model": "claude-haiku-5-5"}
++{"work_id": 4401497158, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "Hybrid inductor-current sensor for a high-frequency BUCK DC-DC converter implemented in CMOS; circuit and low-power hardware design.", "model": "claude-haiku-5-5"}
++{"work_id": 3188260334, "categories": ["compilers"], "primary": "compilers", "confidence": 0.7, "rationale": "Phase-based splitting of lookup caches to avoid monomorphization bottlenecks in managed runtime JIT optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 2992990171, "categories": ["nlp"], "primary": null, "confidence": 0.2, "rationale": "College English teaching reform policy paper; education rather than computer science research, so at most a weak language-related fit with no strong category.", "model": "claude-haiku-5-5"}
++{"work_id": 4415934317, "categories": ["programming_languages", "compilers"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Cloud Sisal language, semantic transformations and specialization of general-purpose programs via reducing concretizations; compiler component present.", "model": "claude-haiku-5-5"}
++{"work_id": 3020776133, "categories": ["compilers", "deep_learning_architectures"], "primary": "compilers", "confidence": 0.55, "rationale": "Translator from NNEF IR to clDNN for executing neural network inference on OpenCL devices; ML compiler/hardware mapping.", "model": "claude-haiku-5-5"}
++{"work_id": 2794424394, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Control theory for memristive neural networks with sampling and saturation; stability analysis framed as control theory, closest fit is ml_theory, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4206295199, "categories": ["computer_vision", "ml_theory"], "primary": "computer_vision", "confidence": 0.4, "rationale": "Adaptive electrical capacitance tomography image reconstruction with relevance vector machine; imaging inverse problem.", "model": "claude-haiku-5-5"}
++{"work_id": 4403183224, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Stability analysis and control of semi-Markov jump linear systems is classical control theory, not a clear CS category in the ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 4380434351, "categories": ["compilers", "computer_architecture"], "primary": "compilers", "confidence": 0.6, "rationale": "Compilation and optimization for heterogeneous multicore embedded platforms, source-to-source compilation.", "model": "claude-haiku-5-5"}
++{"work_id": 3134130217, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational research on digital resources in early childhood education; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2906120018, "categories": ["programming_languages", "software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Static code analyzer for an algorithms education portal using SSA and dataflow analysis to find errors.", "model": "claude-haiku-5-5"}
++{"work_id": 2897836007, "categories": ["compilers"], "primary": "compilers", "confidence": 0.6, "rationale": "Automatic parallelization of loop recurrences as scans and reductions via function reconstruction, a compiler optimization technique.", "model": "claude-haiku-5-5"}
++{"work_id": 3193709667, "categories": ["compilers", "algorithms"], "primary": "compilers", "confidence": 0.5, "rationale": "Generating parallel tiled OpenMP code for linear recurrence equations through loop transformation and wavefront parallelism.", "model": "claude-haiku-5-5"}
++{"work_id": 4392877469, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Fuzzy control applied to a locally built microcontroller-based process controller; embedded hardware and control system implementation, not clearly AI or ML core.", "model": "claude-haiku-5-5"}
++{"work_id": 2981857535, "categories": ["database_systems", "compilers"], "primary": "database_systems", "confidence": 0.8, "rationale": "Machine code caching in a PostgreSQL query JIT compiler; query processing performance with compilation as the mechanism.", "model": "claude-haiku-5-5"}
++{"work_id": 2886921199, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Model-based iterative learning control for wind turbine loads; control theory with reduced-order modelling, weakly fits ML categories, possibly none.", "model": "claude-haiku-5-5"}
++{"work_id": 4226087031, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.45, "rationale": "Array of Things urban wireless sensor network instrument; sensor networks and smart city infrastructure with community engagement.", "model": "claude-haiku-5-5"}
++{"work_id": 4392445319, "categories": ["hci"], "primary": "hci", "confidence": 0.55, "rationale": "Sequence mining of player log data in a puzzle game to assess problem-solving behaviour; educational games and user behaviour analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4416730187, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Development of a school teacher administration application via ADDIE; educational technology rather than computer science research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3138455952, "categories": ["database_systems", "compilers"], "primary": "compilers", "confidence": 0.6, "rationale": "Query compilation targeting GPUs; code generation and optimisation for database queries on parallel hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 2912097352, "categories": ["compilers", "cybersecurity"], "primary": "compilers", "confidence": 0.5, "rationale": "Compiling code to handle hidden deterministic hardware aliasing and encrypted computing; compilation technique with security motivation.", "model": "claude-haiku-5-5"}
++{"work_id": 4384026566, "categories": ["compilers"], "primary": "compilers", "confidence": 0.7, "rationale": "Automatic optimisation of straightline code outperforming compilers; code optimisation pipeline.", "model": "claude-haiku-5-5"}
++{"work_id": 4411259674, "categories": ["compilers", "formal_verification"], "primary": "compilers", "confidence": 0.6, "rationale": "Self-adaptive compilation framework switching between formally verified and optimised compilation for embedded systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7142973430, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Development and validation of assessment instruments for science education; education research, not core CS, so only weakly fits; assigned conservatively to software_engineering as instrument development is closest, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 3161777593, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Vocational education textbook design; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4290851516, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Unsupervised source-domain selection for domain adaptation in remote sensing; methodological domain adaptation with remote sensing application, could also be computer_vision.", "model": "claude-haiku-5-5"}
++{"work_id": 7203891956, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.7, "rationale": "Region-based memory management and tail recursion semantics in ML, a language implementation topic.", "model": "claude-haiku-5-5"}
++{"work_id": 3082269642, "categories": ["compilers"], "primary": "compilers", "confidence": 0.75, "rationale": "Automated transformation of Simulink models to multi-core parallel code via dataflow IR, a compilation/parallelisation pipeline.", "model": "claude-haiku-5-5"}
++{"work_id": 4386632677, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Extensible compact Scheme VM design; language implementation, with compiler-driven specialization.", "model": "claude-haiku-5-5"}
++{"work_id": 7160557760, "categories": ["compilers", "ai_accelerators"], "primary": "compilers", "confidence": 0.8, "rationale": "AOT compiler for PyTorch CPU inference emitting LLVM IR; ML compiler for inference including LLM workloads.", "model": "claude-haiku-5-5"}
++{"work_id": 7131081605, "categories": ["database_systems", "compilers"], "primary": "database_systems", "confidence": 0.6, "rationale": "Query optimizer for DataFrame libraries, a data-management problem with source-to-source rewriting.", "model": "claude-haiku-5-5"}
++{"work_id": 4403493438, "categories": ["compilers"], "primary": "compilers", "confidence": 0.7, "rationale": "Abstract interpretation-based JIT optimizer for CPython generated from a DSL.", "model": "claude-haiku-5-5"}
++{"work_id": 7115276129, "categories": ["compilers"], "primary": "compilers", "confidence": 0.7, "rationale": "Common compiler ecosystem for DSLs targeting high-performance computing.", "model": "claude-haiku-5-5"}
++{"work_id": 2905512450, "categories": ["hci", "speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Audio spatialisation toolkit and hearing loss/hearing aid simulation for virtual environments; spatial audio rendering with HRTF, framed around hearing simulation for users.", "model": "claude-haiku-5-5"}
++{"work_id": 4415312216, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Sound field reproduction and multi-channel equalization for automotive audio, using neural optimization for filter design.", "model": "claude-haiku-5-5"}
++{"work_id": 7219107916, "categories": ["speech_audio", "computer_vision", "hci"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Environmental sound classification via ResNet-50 on spectrogram images, delivered as an assistive Android app for hearing-impaired users.", "model": "claude-haiku-5-5"}
++{"work_id": 2805996599, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Age identification from voice using GMM/DNN and MFCC features, a speaker-characterisation task.", "model": "claude-haiku-5-5"}
++{"work_id": 4388462287, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Chapter on microphone use techniques in audio engineering practice; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4307835756, "categories": ["algorithms", "speech_audio"], "primary": "algorithms", "confidence": 0.5, "rationale": "Sparse array design for DOA estimation of mixed near-field and far-field signals; signal processing with array geometry combinatorics.", "model": "claude-haiku-5-5"}
++{"work_id": 4414880351, "categories": ["speech_audio", "cybersecurity"], "primary": "speech_audio", "confidence": 0.5, "rationale": "Acoustic drone detection using HHT and MFCC features with a lightweight classifier; audio classification applied to UAV detection.", "model": "claude-haiku-5-5"}
++{"work_id": 4384303965, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Audio-visual multi-channel speech separation, dereverberation and recognition with Conformer ASR back-end.", "model": "claude-haiku-5-5"}
++{"work_id": 4385474051, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Pre-training end-to-end ASR models with augmented speech samples generated from unpaired data.", "model": "claude-haiku-5-5"}
++{"work_id": 4386436058, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Contrastive audio-video dual encoder learning multimodal representations for music videos.", "model": "claude-haiku-5-5"}
++{"work_id": 6966747421, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.6, "rationale": "MFCC-based analysis of L2 accented speech with ML classification; speech signal processing applied to L1 transfer, with applied linguistics framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7162373303, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Music source separation comparing DenseNet-based CNN models for vocal/accompaniment separation.", "model": "claude-haiku-5-5"}
++{"work_id": 4387489607, "categories": ["deep_learning_architectures", "speech_audio"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "DOA estimation for array signals using autoencoder and parallel VGG networks; signal processing with a deep architecture, not clearly speech.", "model": "claude-haiku-5-5"}
++{"work_id": 4392902829, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Challenge overview for audio-visual target speaker extraction, a speech enhancement task with visual cues.", "model": "claude-haiku-5-5"}
++{"work_id": 3156731511, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Language model integration for attention-based encoder-decoder speech recognition models.", "model": "claude-haiku-5-5"}
++{"work_id": 7139231366, "categories": ["speech_audio", "computational_biology"], "primary": "speech_audio", "confidence": 0.45, "rationale": "Music performance analysis with DTW on harp recordings for rhythm; music information retrieval, though pedagogical focus.", "model": "claude-haiku-5-5"}
++{"work_id": 7155062195, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Latent-space diffusion-based speech restoration aimed at fast inference.", "model": "claude-haiku-5-5"}
++{"work_id": 2936511587, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Domain adaptation of DNN speaker embeddings using autoencoders for speaker recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 2910678517, "categories": ["speech_audio", "computer_vision"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Geometric calibration of distributed microphone arrays using cross-correlation, with computer vision techniques for pose estimation.", "model": "claude-haiku-5-5"}
++{"work_id": 4412729123, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Music production and mixing guidance; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4400037432, "categories": ["speech_audio", "hci"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Speech-based emotion recognition via bispectral analysis and CNN; affective computing touches hci but primary is speech processing.", "model": "claude-haiku-5-5"}
++{"work_id": 2883033779, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.5, "rationale": "Blind source separation of signals via invasive weed optimization; signal separation mainly applied to audio; optimization-based, could also be algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 2991471437, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Music information retrieval: learning sonata form structure from scores with HMMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4327772893, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.5, "rationale": "Music composition and score recognition with deep learning; music domain, generation aspect.", "model": "claude-haiku-5-5"}
++{"work_id": 7172368152, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Piano acoustics and mechanics study; physics/musical performance rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7166639495, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Target speaker extraction and ASR for multi-party conversations using multichannel audio.", "model": "claude-haiku-5-5"}
++{"work_id": 7169835704, "categories": ["speech_audio", "generative_ai"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Zero-shot music editing with diffusion transformers and rectified flow; generative model contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4411133045, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.65, "rationale": "Multi-label bird sound classification with CNN and LSTM on mel spectrograms.", "model": "claude-haiku-5-5"}
++{"work_id": 7140139406, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Automatic speech recognition transcription system for Irish, with diarisation and punctuation restoration.", "model": "claude-haiku-5-5"}
++{"work_id": 4411733613, "categories": ["speech_audio", "hci"], "primary": "hci", "confidence": 0.55, "rationale": "Evaluation of ASR-based reading tutor feedback effects on children; user study framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4405633169, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Continual learning for ASR using machine speech chain and GEM; primarily speech recognition.", "model": "claude-haiku-5-5"}
++{"work_id": 4313196217, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Audio classification of Vietnamese folk songs via CNN; music information retrieval.", "model": "claude-haiku-5-5"}
++{"work_id": 3197227964, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Personalized speech enhancement with self-supervised data purification.", "model": "claude-haiku-5-5"}
++{"work_id": 2932055495, "categories": ["speech_audio", "computer_vision"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Binaural sound source localisation from waveform with CNN; audio signal processing.", "model": "claude-haiku-5-5"}
++{"work_id": 7122349630, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Systematic survey of automatic music plagiarism detection, a music information retrieval topic.", "model": "claude-haiku-5-5"}
++{"work_id": 2980513516, "categories": ["speech_audio", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Countermeasure protecting speaker verification from spoofing attacks; security framing with speech features.", "model": "claude-haiku-5-5"}
++{"work_id": 2965346518, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Cross-modal representation of sound with colour and shape in musical scores; notation and visualisation design.", "model": "claude-haiku-5-5"}
++{"work_id": 3094199623, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.75, "rationale": "Real-time time-domain speech enhancement using a dual-path self-attention RNN architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 3196692235, "categories": ["speech_audio", "computational_biology"], "primary": "speech_audio", "confidence": 0.6, "rationale": "COVID-19 detection from cough audio using acoustic cues; health-focused audio analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4386556071, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Self-supervised disentanglement of rhythm and harmony in music audio with a VAE; music generation.", "model": "claude-haiku-5-5"}
++{"work_id": 4390905911, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Streaming ASR domain adaptation for Japanese speech using density ratio approach and CTC decoder; speech recognition is primary.", "model": "claude-haiku-5-5"}
++{"work_id": 2982577960, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Single-channel blind source separation of audio/speech signals using EMD and NMF.", "model": "claude-haiku-5-5"}
++{"work_id": 3160275121, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.65, "rationale": "Multilingual phonetic dataset for phone recognition in low-resource speech.", "model": "claude-haiku-5-5"}
++{"work_id": 4212954326, "categories": ["computer_architecture"], "primary": null, "confidence": 0.2, "rationale": "Sparse array beamforming for sonar/underwater signal processing; not clearly in any ontology category, closest is signal processing with weak fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3134129880, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Magazine piece on smart speaker acoustic sensors; a general technology overview with weak fit to speech_audio or hci.", "model": "claude-haiku-5-5"}
++{"work_id": 7156277776, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.55, "rationale": "Acoustic drone detection and localization with a compact CNN on audio; audio event detection with TinyML deployment.", "model": "claude-haiku-5-5"}
++{"work_id": 7148385849, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Self-supervised speech representations evaluated on speaker diarization and separation.", "model": "claude-haiku-5-5"}
++{"work_id": 3033627755, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.7, "rationale": "End-to-end neural speaker diarization with speaker-wise chain rule for variable speaker counts.", "model": "claude-haiku-5-5"}
++{"work_id": 4415338423, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Benchmarking of speaker diarization models across datasets and languages.", "model": "claude-haiku-5-5"}
++{"work_id": 4411173190, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.55, "rationale": "Word classification from surface EMG speech signals for unvoiced speech recognition using classical ML and dimensionality reduction.", "model": "claude-haiku-5-5"}
++{"work_id": 7105655369, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Scoping review protocol on knowledge graphs and semantic technologies in sport science; knowledge graphs are data-management artifacts but the domain is sport/exercise science, so only weakly fits computing categories.", "model": "claude-haiku-5-5"}
++{"work_id": 4412742458, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Systematic review of AI-generated instructional videos (AIGC, GANs, large models) and their learning impacts, touching on educational human factors.", "model": "claude-haiku-5-5"}
++{"work_id": 3097725556, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Process mining connector extracting event logs from the Camunda workflow engine; event data extraction and mining fit data management, with BPM context.", "model": "claude-haiku-5-5"}
++{"work_id": 4317502683, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.45, "rationale": "Data integration of hospital and management information systems and its effect on report accuracy; health informatics framed as data integration.", "model": "claude-haiku-5-5"}
++{"work_id": 4411208804, "categories": ["distributed_computing", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Scalable AI-CRM architecture with microservices, Kubernetes, serverless and cloud infrastructure; cloud-native design is the main focus.", "model": "claude-haiku-5-5"}
++{"work_id": 7107958642, "categories": ["programming_languages", "database_systems"], "primary": "programming_languages", "confidence": 0.55, "rationale": "Course on Datalog and its extensions, rule languages and semantics, with a modern Datalog engine.", "model": "claude-haiku-5-5"}
++{"work_id": 2991196924, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Integrating Linked Open Data into a data cube for organisational value; semantic data integration, though abstract is minimal.", "model": "claude-haiku-5-5"}
++{"work_id": 4414507907, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.8, "rationale": "Hybrid SQL/NoSQL database design with AI-driven optimisations; clearly a database systems contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4225908445, "categories": ["database_systems", "algorithms"], "primary": "database_systems", "confidence": 0.6, "rationale": "Skyline query computation on massive incomplete data; query processing over data tables with algorithmic design.", "model": "claude-haiku-5-5"}
++{"work_id": 4307637929, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Human-machine collaboration and human-AI interaction in strategic marketing decisions; a survey-based business perspective on human-AI interaction.", "model": "claude-haiku-5-5"}
++{"work_id": 7218431496, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Data mining with C4.5 for sales prediction; data mining as a data-management problem, classification-based forecasting.", "model": "claude-haiku-5-5"}
++{"work_id": 3018396859, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.7, "rationale": "Experimental study of Hive query performance across execution engines on SQL-on-Hadoop systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2604259521, "categories": ["database_systems", "nlp"], "primary": "database_systems", "confidence": 0.5, "rationale": "Knowledge base construction from richly formatted data; relation extraction with a data model and system design.", "model": "claude-haiku-5-5"}
++{"work_id": 2920557606, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Hyperspectral imaging of the iris; instrument performance for image acquisition, though biomedical in nature.", "model": "claude-haiku-5-5"}
++{"work_id": 2951532316, "categories": ["database_systems", "robotics"], "primary": "database_systems", "confidence": 0.6, "rationale": "Database extension for spatio-temporal stream management and continuous queries on AUVs.", "model": "claude-haiku-5-5"}
++{"work_id": 4327772251, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "General data mining design and implementation in information management systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2799038541, "categories": ["database_systems", "computational_complexity"], "primary": "database_systems", "confidence": 0.6, "rationale": "Index structures for enumerating MSO queries over strings under updates; query evaluation with update efficiency.", "model": "claude-haiku-5-5"}
++{"work_id": 4289533871, "categories": ["database_systems", "algorithms", "distributed_computing"], "primary": "database_systems", "confidence": 0.6, "rationale": "Subgraph matching in distributed graph systems; query processing over large graphs.", "model": "claude-haiku-5-5"}
++{"work_id": 2942479669, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.6, "rationale": "Converting unstructured agro-data to structured form using a NoSQL document database.", "model": "claude-haiku-5-5"}
++{"work_id": 7131351419, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "AI-assisted counselor workflow system evaluated through practical use; closer to human-centred system design than a clear core category.", "model": "claude-haiku-5-5"}
++{"work_id": 7159658581, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Applies a Naive Bayes classifier as a decision tool for hiring; a probabilistic classification method application, weakly fitting ml_theory, with no clear fit to other categories.", "model": "claude-haiku-5-5"}
++{"work_id": 4401268941, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Library science textbook on data librarianship; not primarily a computer-science research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7161581487, "categories": ["database_systems", "information_retrieval"], "primary": "database_systems", "confidence": 0.7, "rationale": "Filtered approximate nearest neighbor search with hybrid vector-attribute queries; query processing over databases, with retrieval context for RAG.", "model": "claude-haiku-5-5"}
++{"work_id": 4387233410, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Applies neural network with particle swarm optimization to student graduation prediction; a generic data-mining application with weak fit to any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 2997642018, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Developing a payroll information system for a laundry company; an applied business information system rather than a CS research category, closest to software_engineering but weak.", "model": "claude-haiku-5-5"}
++{"work_id": 3168932797, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Record linkage and data integration across health databases; matches data integration and cleaning in database_systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7118687013, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.4, "rationale": "Systematic review of generative AI and deep learning in personalized education; applied education context, weakly tied to llm and nlp.", "model": "claude-haiku-5-5"}
++{"work_id": 4401822397, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Applies K-Nearest Neighbor classification to predict student graduation; generic classification application with weak category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3134393417, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Studies fault detection and mitigation in data pipelines via action research; software and data engineering quality.", "model": "claude-haiku-5-5"}
++{"work_id": 3201407754, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Compares regression, neural network and SVM to predict vocational student employability; generic data-mining application with weak category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4205179238, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Library book-lending information system design for a madrasah; applied information system development, closest to interactive system design but weak fit overall.", "model": "claude-haiku-5-5"}
++{"work_id": 4416374389, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Web system implementation for sales and inventory processes evaluated quantitatively; business information system development, weakly fits software engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 4255118578, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "SQL Server wait statistics and performance troubleshooting from a book chapter; database performance topic.", "model": "claude-haiku-5-5"}
++{"work_id": 7160618846, "categories": ["llm", "distributed_computing"], "primary": "llm", "confidence": 0.6, "rationale": "Observability platform for LLM applications with storage and benchmarking; LLM-focused systems work with database-like storage design.", "model": "claude-haiku-5-5"}
++{"work_id": 2982402713, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.85, "rationale": "Garbage collection of MVCC versions in in-memory HTAP database systems; transaction processing.", "model": "claude-haiku-5-5"}
++{"work_id": 6944527151, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.85, "rationale": "Analytical database system features, query planning cost model and Spark integration in Actian Vector.", "model": "claude-haiku-5-5"}
++{"work_id": 3036960336, "categories": ["hci"], "primary": "hci", "confidence": 0.45, "rationale": "Dashboard usage and feedback for general practice clinicians; user-facing interactive system evaluation in health informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 2933476764, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.6, "rationale": "Software-defined database cluster replicating data across heterogeneous DBMSs.", "model": "claude-haiku-5-5"}
++{"work_id": 3008451085, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.8, "rationale": "Highly available HTAP distributed DBMS with consensus-based serialization for IoT.", "model": "claude-haiku-5-5"}
++{"work_id": 4402761012, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Portal aggregating paleontological specimen databases; data integration of museum records, though domain is biodiversity informatics.", "model": "claude-haiku-5-5"}
++{"work_id": 7116397644, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "K-Means clustering of drug sales data for inventory management; data mining as a data-management problem, applied to inventory, with weak fit to any ML category.", "model": "claude-haiku-5-5"}
++{"work_id": 3189460411, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Management/strategy study of firm coopetition networks using social network analysis; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3022939691, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Complex event stream processing and forecasting with latent representations; closest to data management of event streams, though it draws on ML methods.", "model": "claude-haiku-5-5"}
++{"work_id": 3010181845, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.3, "rationale": "Comparison of Naive Bayes and KNN for student graduation classification via data mining; conservative fit is weak, arguably no clear CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4282592169, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.7, "rationale": "Learning normalized database schemas from denormalized tables; a data management problem solved with unsupervised learning.", "model": "claude-haiku-5-5"}
++{"work_id": 3043717349, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Formal translation and combination of belief representations (probability, fuzzy, Dempster-Shafer) for uncertainty reasoning; closest to probabilistic modelling theory.", "model": "claude-haiku-5-5"}
++{"work_id": 7159820960, "categories": ["database_systems", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.75, "rationale": "Analysis of integrating relational databases with cloud environments, architectures, and providers.", "model": "claude-haiku-5-5"}
++{"work_id": 4220806123, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.45, "rationale": "Association rule mining (Apriori with a three-dimensional matrix) on equipment decommissioning data; data mining with a power-systems application.", "model": "claude-haiku-5-5"}
++{"work_id": 3183141872, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Market basket association rule mining and frequent itemset mining on purchase histories; data mining as a data-management problem.", "model": "claude-haiku-5-5"}
++{"work_id": 7212066750, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Development of a web fitness management system evaluated by usability heuristics and user satisfaction; closest to usability/interactive systems, though it is largely a web-application development paper.", "model": "claude-haiku-5-5"}
++{"work_id": 4416057735, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.8, "rationale": "Dependent session type theory extending TLL with concurrency; type system and semantics are the core, with verification of programs as motivation.", "model": "claude-haiku-5-5"}
++{"work_id": 7155176289, "categories": ["cybersecurity", "programming_languages"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Binary format metadata for disassembly and lifting supports binary security analysis and instrumentation; also touches compilation and binary representation.", "model": "claude-haiku-5-5"}
++{"work_id": 4384069307, "categories": ["hci"], "primary": "hci", "confidence": 0.45, "rationale": "Collaborative programming system for students with a quasi-experimental user study of learning and group metacognition; educational technology framing, not clearly a core CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4388890939, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.75, "rationale": "Analysis of structural typing support and extensible records in Scala 3 type system design.", "model": "claude-haiku-5-5"}
++{"work_id": 4386082988, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Static analysis via abstract interpretation for verification of GRAFCET specifications in industrial control.", "model": "claude-haiku-5-5"}
++{"work_id": 2942810970, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Educational prototype of virus behavior for understanding and prevention of malware.", "model": "claude-haiku-5-5"}
++{"work_id": 7161825823, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Applies machine learning with spatial sample weighting to land cover change modeling; primarily a geographic application, so weakly fits the ontology; ml_theory is the closest methodological fit but low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4390043981, "categories": ["algorithms", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.45, "rationale": "Robust approximate dynamic programming for large-scale power unit commitment; optimization method applied to energy systems, with algorithmic contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4412097039, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Python-based computer visualization of physical phenomena in education, modeling falling body with drag; computational modeling tooling with educational framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7152032912, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Thesis on formal methods, Petri nets and model checking to verify robust behavior of autonomous robotic systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4293812983, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.9, "rationale": "Gradual dependently typed language design with propositional equality, type safety and metatheory; core is type theory and language semantics.", "model": "claude-haiku-5-5"}
++{"work_id": 3046578987, "categories": ["computational_complexity", "formal_verification"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Process calculus with stochastic bisimulation and performance evaluation of Petri nets; closest to formal methods/verification of concurrent systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4394658783, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Categorical semantics of causal morphisms, coinduction and recursion in traced monoidal categories; theoretical semantics of computation.", "model": "claude-haiku-5-5"}
++{"work_id": 7128354711, "categories": ["software_engineering", "database_systems"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Waterfall-based design and implementation of a web inventory information system with black-box testing.", "model": "claude-haiku-5-5"}
++{"work_id": 4296940370, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Concluding chapter of a Haskell book covering functors, monads, GADTs and type systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3091828950, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Research model in didactics of programming grounded in Piaget's epistemology; computer science education rather than CS technical research.", "model": "claude-haiku-5-5"}
++{"work_id": 7124299435, "categories": ["ai_agents", "formal_verification"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Enactability of multi-agent interaction protocols specified with trace expressions, a multi-agent systems topic with formal aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 4320558734, "categories": ["software_engineering", "programming_languages"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Tool paper providing IDE language support (syntax highlighting, autocompletion, feedback) for the COOL teaching language.", "model": "claude-haiku-5-5"}
++{"work_id": 4414483419, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Errata to a mathematical paper on models of mixed choice in concurrency/process semantics; theoretical and largely mathematical.", "model": "claude-haiku-5-5"}
++{"work_id": 7128043521, "categories": ["ml_theory", "programming_languages"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Proof-theoretic account of Bayesian networks via linear logic proof nets; probabilistic models framed formally.", "model": "claude-haiku-5-5"}
++{"work_id": 4392822342, "categories": ["programming_languages", "formal_verification"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Proves correctness and stability of mergesort in Rocq using relational parametricity; mainly verification of programs with a PL-semantics flavour.", "model": "claude-haiku-5-5"}
++{"work_id": 4367318871, "categories": ["programming_languages", "compilers"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Studies WebAssembly versus JavaScript execution performance, touching language implementation and execution of code in browsers.", "model": "claude-haiku-5-5"}
++{"work_id": 3194305803, "categories": ["software_engineering", "programming_languages"], "primary": "software_engineering", "confidence": 0.7, "rationale": "A library and DSL for property-based testing models of stateful programs, a software testing contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4206497465, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Book chapter with no substantive abstract; insufficient information to assign a computer-science category confidently.", "model": "claude-haiku-5-5"}
++{"work_id": 4320928251, "categories": ["software_engineering", "programming_languages"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Contract-based design and type constraints for specifying extra-functional properties in component design.", "model": "claude-haiku-5-5"}
++{"work_id": 7164912132, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.7, "rationale": "Synthesis of best abstract transformers for static analysis via optimization modulo theories, a verification/automated reasoning task.", "model": "claude-haiku-5-5"}
++{"work_id": 2969568901, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.85, "rationale": "Type system with linear and non-linear recursive types and categorical semantics, core programming language theory.", "model": "claude-haiku-5-5"}
++{"work_id": 7140223191, "categories": ["algorithms", "information_retrieval"], "primary": "algorithms", "confidence": 0.4, "rationale": "Canonical string representation of expression DAGs for symbolic regression search; algorithmic representation with ML application, uncertain fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3204449783, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Student awareness evaluation of an e-learning course, a user study of an interactive educational system.", "model": "claude-haiku-5-5"}
++{"work_id": 7220384869, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Development of a web e-commerce application for a small business with system testing; applied web system development.", "model": "claude-haiku-5-5"}
++{"work_id": 2955161317, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.4, "rationale": "Chapter on recognizable tree languages and algebraic properties; closest to automata/theory of tree languages, which falls under computability and automata theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4307886855, "categories": ["programming_languages", "software_engineering"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Derives incremental type checkers from Statix type system specifications using scope graphs; type systems and language tooling, with IDE incrementality aspect.", "model": "claude-haiku-5-5"}
++{"work_id": 2892794698, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Development of an information system for programming courses using waterfall method; a web/app system build, weakly fits software engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 4413180303, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.35, "rationale": "Power flow visualization and disturbance prediction in smart grids via graphs; power system analysis tooling, loosely scientific/engineering computing.", "model": "claude-haiku-5-5"}
++{"work_id": 6891551213, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.3, "rationale": "Artifact for a paper on shape analysis with gradual types; abstract is minimal, topic inferred from title suggests program analysis and type systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2990464075, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Plenary talk on Industry 4.0/5.0 and cyber-physical-social systems and enterprise management; not a clear computer-science research contribution in a specific category.", "model": "claude-haiku-5-5"}
++{"work_id": 3199165787, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Practice report on a web-based alumni tracer information system; information-system development with no clear ontology category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3010153186, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Variational Bayes nonlinear state estimation filter with probabilistic approximation; statistical estimation methodology, applied to target tracking.", "model": "claude-haiku-5-5"}
++{"work_id": 3030633439, "categories": [], "primary": null, "confidence": 0.25, "rationale": "Essay on the value of smart education and big data; education policy discussion rather than computer-science research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 2941342584, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Android academic information system aimed at supporting students' study activities and smartphone use; closest to interactive user-facing systems, weak fit.", "model": "claude-haiku-5-5"}
++{"work_id": 7215646023, "categories": ["cybersecurity", "computational_biology"], "primary": null, "confidence": 0.2, "rationale": "Community service report on an Android expert system for orchid disease diagnosis; application domain is plant pathology rather than a clear CS category, so only a weak fit to expert systems; conservative assignment.", "model": "claude-haiku-5-5"}
++{"work_id": 4287283171, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Epistemic separation logic with action models is a formal logic for reasoning about knowledge and information change; closest to ml_theory-adjacent formal logic, but it is really logic/formal semantics, so low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4388795790, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Community training on App Inventor for high school students; educational outreach rather than a computer-science research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4386211482, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Proceedings of logic programming conference; the abstract describes abstract interpretation domains for static analysis of logic programs, a program-analysis-as-language problem.", "model": "claude-haiku-5-5"}
++{"work_id": 6910567405, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Software release of PtrHash, a minimal perfect hash function implementation, which is a data-structure/algorithm artifact.", "model": "claude-haiku-5-5"}
++{"work_id": 4281724929, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.4, "rationale": "Implicit state machines for real-time and reactive embedded systems; programming abstraction for FSMs, with compiler/language flavour but abstract is brief.", "model": "claude-haiku-5-5"}
++{"work_id": 3121979527, "categories": ["programming_languages", "formal_verification"], "primary": "formal_verification", "confidence": 0.7, "rationale": "Decentralized analysis ensuring protocol conformance and deadlock-freedom of multiparty session-typed processes; type-system-based verification of concurrent protocols.", "model": "claude-haiku-5-5"}
++{"work_id": 7153697985, "categories": ["programming_languages", "cybersecurity"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Dissertation on information-flow control and timing-sensitive security in runtime systems and language design, with Coq proofs; primarily language-based security.", "model": "claude-haiku-5-5"}
++{"work_id": 2810401719, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Cubical type theory foundations for informal homotopy type theory, a proof-assistant style formalisation of higher groupoid structure.", "model": "claude-haiku-5-5"}
++{"work_id": 2888336873, "categories": ["compilers", "computer_architecture"], "primary": "compilers", "confidence": 0.5, "rationale": "Static worst-case execution time estimation of IEC 61131-3 control code via abstract interpretation for embedded automation systems; closer to program analysis/embedded timing, with compiler-style analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 3009620545, "categories": ["blockchain", "database_systems"], "primary": "blockchain", "confidence": 0.8, "rationale": "Hyperledger Fabric-based maintenance documentation system; blockchain application with performance evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 4416148866, "categories": ["blockchain", "applied_cryptography"], "primary": "blockchain", "confidence": 0.8, "rationale": "Blockchain lending transactions with weighted threshold signature scheme; cryptographic primitive is a key contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4401513854, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Review of blockchain for supply chain transparency; blockchain applications.", "model": "claude-haiku-5-5"}
++{"work_id": 4402990505, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Consumer perspective on blockchain; primarily a business/consumer study rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4401788847, "categories": ["blockchain", "edge_computing"], "primary": "blockchain", "confidence": 0.6, "rationale": "Smart gateways collecting endpoint data and uploading to blockchain; IoT data collection at the edge.", "model": "claude-haiku-5-5"}
++{"work_id": 2977293541, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Introductory German book on blockchain basics for business; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7127122082, "categories": ["blockchain", "nlp"], "primary": "blockchain", "confidence": 0.7, "rationale": "Converting natural-language legal contracts into smart contract code; smart contract generation with NLP parsing.", "model": "claude-haiku-5-5"}
++{"work_id": 4402966691, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Decentralized charity platform built on smart contracts and DeFi mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 7152727986, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Overview of blockchain applications for healthcare data security.", "model": "claude-haiku-5-5"}
++{"work_id": 2989399708, "categories": ["blockchain", "edge_computing"], "primary": "blockchain", "confidence": 0.6, "rationale": "Smart-contract payments for platooning in vehicular IoT; blockchain-enabled vehicular system.", "model": "claude-haiku-5-5"}
++{"work_id": 7214298050, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Public opinion study on cryptocurrency regulation in criminal justice/policy venue; not a computer-science contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3088531038, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.9, "rationale": "Ethereum-based decentralized prototype for healthcare data integrity.", "model": "claude-haiku-5-5"}
++{"work_id": 3014880043, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Data mining (DBSCAN anomaly detection) on Bitcoin price data; closest fit is data mining as data-management problem, though it is weakly CS-core.", "model": "claude-haiku-5-5"}
++{"work_id": 4400034579, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Blockchain-based e-vault system for legal records using smart contracts.", "model": "claude-haiku-5-5"}
++{"work_id": 4416430007, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.5, "rationale": "Data sharing via Eclipse Dataspace Connector for Digital Product Passports; data-space infrastructure rather than blockchain, so low-confidence fit to distributed/data management; blockchain is weak here.", "model": "claude-haiku-5-5"}
++{"work_id": 4400485117, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.75, "rationale": "Framework integrating digital twins and private blockchain for energy performance contracts.", "model": "claude-haiku-5-5"}
++{"work_id": 2901943454, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Systematic literature review of blockchain use cases for smart cities.", "model": "claude-haiku-5-5"}
++{"work_id": 3194765017, "categories": [], "primary": null, "confidence": 0.9, "rationale": "News item about an NFT auction; not computer-science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4382770016, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Empirical study of blockchain effects on corporate governance; business/management focus with blockchain as the technology studied.", "model": "claude-haiku-5-5"}
++{"work_id": 4407826131, "categories": ["blockchain", "deep_learning_architectures"], "primary": "blockchain", "confidence": 0.6, "rationale": "Survey on integrating blockchain with neural networks for predictive modeling; blockchain is the primary focus with ANNs secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 4245301215, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Business book on decentralization and blockchain strategy; not a clear CS research contribution, so no category assigned conservatively.", "model": "claude-haiku-5-5"}
++{"work_id": 3143290424, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Management-oriented literature review of blockchain research trends in business; not a CS technical contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7125399881, "categories": ["blockchain", "generative_ai", "cybersecurity"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain-anchored media authentication to counter deepfakes; blockchain is the main system contribution, with deepfake/AI-generated content detection and cybersecurity framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7117461924, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Consumer behavior study using SEM on green payments with blockchain and AI cues; primarily business/marketing research rather than CS contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7164945802, "categories": ["federated_learning", "cybersecurity", "blockchain"], "primary": "federated_learning", "confidence": 0.75, "rationale": "Federated learning with XAI for fraud detection on blockchain transactions; federated learning is the core method, with fraud detection and blockchain as application context.", "model": "claude-haiku-5-5"}
++{"work_id": 3103101544, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Permissioned blockchain-based local P2P energy market with privacy-by-design; blockchain systems are the central contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3012297310, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain-based pet medical data sharing model with improved consensus; blockchain system design.", "model": "claude-haiku-5-5"}
++{"work_id": 4417496166, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Survey-based business study of blockchain adoption and consumer trust in SMEs; management research rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4327927520, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.75, "rationale": "Blockchain-based health record management system across hospitals; blockchain system design applied to healthcare.", "model": "claude-haiku-5-5"}
++{"work_id": 4411360648, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Bibliometric topic modeling of Korean blockchain research trends; a scientometric study, not a CS technical contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7160104539, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Chapter on blockchain applied to supply chain management; primary topic is blockchain systems applications.", "model": "claude-haiku-5-5"}
++{"work_id": 4394910888, "categories": ["blockchain", "privacy_preserving_computation"], "primary": "blockchain", "confidence": 0.6, "rationale": "Privacy and security of blockchain systems; blockchain is central with privacy-preserving techniques.", "model": "claude-haiku-5-5"}
++{"work_id": 7160380482, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Blockchain-based healthcare framework with smart contracts and IPFS for EMR and scheduling.", "model": "claude-haiku-5-5"}
++{"work_id": 3095954801, "categories": ["cybersecurity", "computer_networks"], "primary": "computer_networks", "confidence": 0.55, "rationale": "Thesis on secure time synchronization protocols (NTP) and their security analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 4400422781, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.4, "rationale": "Economic analysis of crypto-currency market and blockchain use in finance; mainly economics but touches blockchain systems.", "model": "claude-haiku-5-5"}
++{"work_id": 7143345003, "categories": ["blockchain", "edge_computing"], "primary": "blockchain", "confidence": 0.5, "rationale": "Convergence of IoT and blockchain for secure data sharing.", "model": "claude-haiku-5-5"}
++{"work_id": 4392408184, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain platform for copyright traceability built on FISCO BCOS.", "model": "claude-haiku-5-5"}
++{"work_id": 4386500200, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Survey study on public perception of money and cryptocurrency; behavioral finance, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4210686246, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Policy and geopolitics of digital currencies; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4285316846, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Monetary policy and legal analysis of cryptocurrency payments in Indonesia; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3173158936, "categories": ["blockchain", "cybersecurity"], "primary": "blockchain", "confidence": 0.6, "rationale": "Blockchain framework for pharmaceutical supply chain governance with a cybersecurity maturity model; primarily blockchain applications.", "model": "claude-haiku-5-5"}
++{"work_id": 4206450404, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.8, "rationale": "Blockchain-based energy trading with new consensus mechanism and fairness mechanism.", "model": "claude-haiku-5-5"}
++{"work_id": 4366000231, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Blockchain applied to securing e-learning data; blockchain application chapter.", "model": "claude-haiku-5-5"}
++{"work_id": 7134197021, "categories": ["blockchain", "multimodal_ai", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.4, "rationale": "Multimodal dataset of NFT images, metadata and trading data for pricing analysis; NFT domain is blockchain-related but the dataset is multimodal.", "model": "claude-haiku-5-5"}
++{"work_id": 4379472050, "categories": ["blockchain", "cloud_computing"], "primary": "blockchain", "confidence": 0.6, "rationale": "Blockchain-based architecture for supply chain in cloud with optimisation evaluation.", "model": "claude-haiku-5-5"}
++{"work_id": 3198186069, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Journal dossier introduction in education/language teaching; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4298129476, "categories": ["blockchain", "distributed_computing"], "primary": "blockchain", "confidence": 0.7, "rationale": "Performance evaluation of Ethereum transactions across distributed nodes.", "model": "claude-haiku-5-5"}
++{"work_id": 4315572972, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Blockchain-based information system for verifying educational documents.", "model": "claude-haiku-5-5"}
++{"work_id": 4385250044, "categories": ["blockchain"], "primary": null, "confidence": 0.3, "rationale": "Scientometric analysis of fintech and carbon accounting; bibliometric study, weakly CS-related, blockchain is one topic.", "model": "claude-haiku-5-5"}
++{"work_id": 7123350921, "categories": ["blockchain", "database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Centralised unified double-entry ledger with idempotent webhook processing and SSE; explicitly non-blockchain, closer to data systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3186102448, "categories": ["cloud_computing", "nlp"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Cloud computing applied to e-learning systems in education; primarily cloud service building, with education framing rather than NLP.", "model": "claude-haiku-5-5"}
++{"work_id": 3165257421, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.45, "rationale": "Intelligent temperature monitoring and management in datacenters; datacenter resource management fits cloud/sustainable infrastructure, somewhat ambiguous.", "model": "claude-haiku-5-5"}
++{"work_id": 4296044778, "categories": ["cloud_computing", "sustainable_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Survey of sustainable cloud data center capacity planning.", "model": "claude-haiku-5-5"}
++{"work_id": 2899546521, "categories": ["cloud_computing", "computer_networks"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Rehoming VNF service chains in private network clouds, model-based approach on OpenStack.", "model": "claude-haiku-5-5"}
++{"work_id": 4211066124, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Program file placement in hierarchical cloud-edge platform with reconfiguration control.", "model": "claude-haiku-5-5"}
++{"work_id": 4407737588, "categories": ["cloud_computing", "hci"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Systematic review of mobile cloud gaming's effect on player experience and accessibility; user experience focus suggests hci as well.", "model": "claude-haiku-5-5"}
++{"work_id": 4412993992, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.55, "rationale": "Cloud-native architectures for financial management in food systems; book chapter on cloud-native platforms.", "model": "claude-haiku-5-5"}
++{"work_id": 3125267005, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Bat-algorithm-based task allocation for cloud infrastructure, resource allocation in cloud.", "model": "claude-haiku-5-5"}
++{"work_id": 2794672104, "categories": ["cloud_computing", "sustainable_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Game-theoretic workload exchange among green-energy data centers to reduce energy costs.", "model": "claude-haiku-5-5"}
++{"work_id": 4410342350, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Profit equalization and heuristic resource allocation in cloud federations.", "model": "claude-haiku-5-5"}
++{"work_id": 4409764799, "categories": ["software_engineering", "cloud_computing", "information_retrieval"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Describes design and development of a cloud-based recruitment system (web app architecture, security measures); primarily a system development study with cloud deployment on AWS, ranking component secondary.", "model": "claude-haiku-5-5"}
++{"work_id": 3212132676, "categories": ["distributed_computing", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Service composition with QoS constraints and NSGA-II optimisation over web services; service-oriented computing fits cloud/distributed service selection.", "model": "claude-haiku-5-5"}
++{"work_id": 2980085570, "categories": ["cloud_computing", "hci"], "primary": "cloud_computing", "confidence": 0.4, "rationale": "Evaluates cloud-based group work tools for students and small businesses; cloud-service focus with collaboration tool usability angle.", "model": "claude-haiku-5-5"}
++{"work_id": 3010373421, "categories": ["hci"], "primary": "hci", "confidence": 0.25, "rationale": "Interview study of organizational roles in inter-organizational smart service systems; socio-technical study rather than a core CS technical category, weakly service/IoT engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 4396789527, "categories": ["cybersecurity", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.45, "rationale": "Literature review of security and privacy factors affecting cloud adoption by SMEs; combines cloud adoption with security concerns.", "model": "claude-haiku-5-5"}
++{"work_id": 2912717503, "categories": ["cloud_computing", "edge_computing"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Survey of decentralised cloud models linking cloud with mobile cloud, ad hoc and edge computing.", "model": "claude-haiku-5-5"}
++{"work_id": 809949967, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Distributed coordination and configuration of workflows over web services and handhelds, a distributed systems and service composition prototype.", "model": "claude-haiku-5-5"}
++{"work_id": 7130725155, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Survey of risk assessment, anomaly detection and identity access management for cloud security.", "model": "claude-haiku-5-5"}
++{"work_id": 4400263921, "categories": ["cloud_computing", "cybersecurity"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Simulation of cloud environments for connected vehicles to support anomaly detection; cloud operations monitoring focus.", "model": "claude-haiku-5-5"}
++{"work_id": 3036390872, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Provenance-based audit trail framework for assessing SaaS provider efficiency in cloud environments.", "model": "claude-haiku-5-5"}
++{"work_id": 4391093673, "categories": ["database_systems", "distributed_computing"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Educational big data decision support system with resource scheduling and container-based deployment; mixed framing, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4407956345, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education quality evaluation using AHP and big data policy discussion; not clearly a CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 2796225596, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "SLA translation and resource allocation for service composition, a cloud/service-computing topic.", "model": "claude-haiku-5-5"}
++{"work_id": 7117768918, "categories": ["edge_computing", "distributed_computing"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Empirical study of timeout-based fault detection on edge devices under resource stress.", "model": "claude-haiku-5-5"}
++{"work_id": 4401791482, "categories": ["cloud_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "ML-based security framework for public cloud data protection.", "model": "claude-haiku-5-5"}
++{"work_id": 7208727987, "categories": ["distributed_computing", "software_engineering"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Microservices architecture vs monolith evaluation with message broker for a case-reporting system.", "model": "claude-haiku-5-5"}
++{"work_id": 3087979613, "categories": ["edge_computing", "cloud_computing"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Review of the transition from cloud to fog computing.", "model": "claude-haiku-5-5"}
++{"work_id": 2893549539, "categories": ["cloud_computing", "distributed_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "ML-driven energy-aware resource provisioning for in-memory analytics in cloud platforms.", "model": "claude-haiku-5-5"}
++{"work_id": 3103744586, "categories": ["cloud_computing", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Deploying astronomical image processing pipelines on AWS for LSST data at scale.", "model": "claude-haiku-5-5"}
++{"work_id": 2967724094, "categories": ["cloud_computing", "database_systems"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Real-time VaR and Expected Shortfall computation on a cloud big data analytics platform.", "model": "claude-haiku-5-5"}
++{"work_id": 4405846634, "categories": ["distributed_computing", "cloud_computing", "ml_theory"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Fuzzy clustering for VM resource allocation in cloud data centers; cloud resource management is the main object, with clustering method as technique.", "model": "claude-haiku-5-5"}
++{"work_id": 3043719610, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.9, "rationale": "Evaluates serverless/FaaS versus monolith architectures on Azure, cold starts and scaling.", "model": "claude-haiku-5-5"}
++{"work_id": 3199597950, "categories": ["cloud_computing", "cybersecurity"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Survey of cloud computing issues, service models and security challenges.", "model": "claude-haiku-5-5"}
++{"work_id": 4413436783, "categories": ["cloud_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Security threats and model for multi-tenant Kubernetes clusters, container isolation and secrets management.", "model": "claude-haiku-5-5"}
++{"work_id": 2967650664, "categories": ["cloud_computing", "edge_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Migrating industrial control software to IaaS cloud with real-time latency tests.", "model": "claude-haiku-5-5"}
++{"work_id": 4248407127, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Book chapter on PowerShell data areas; not computer-science research contribution in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 2938024881, "categories": ["sustainable_computing", "distributed_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Data center planning powered by on-site renewable energy with multi-objective negotiation between power and server consumption.", "model": "claude-haiku-5-5"}
++{"work_id": 4243812454, "categories": ["cloud_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Reviews cloud computing security threats and countermeasures for data security.", "model": "claude-haiku-5-5"}
++{"work_id": 4400232270, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.9, "rationale": "Cryptojacking attack vector abusing AWS services, attack analysis and mitigation.", "model": "claude-haiku-5-5"}
++{"work_id": 3094723118, "categories": ["distributed_computing", "reinforcement_learning", "ml_theory"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Threshold load balancing across server pools with asymptotic optimality and online learning of threshold; load balancing in large service systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4407606154, "categories": ["cloud_computing", "cybersecurity"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Monitoring and logging visibility in cloud environments, with SIEM and log encryption elements; primarily cloud operations.", "model": "claude-haiku-5-5"}
++{"work_id": 4401508590, "categories": ["edge_computing", "operating_systems"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Container cold-start acceleration and task scheduling for edge native applications.", "model": "claude-haiku-5-5"}
++{"work_id": 4389723967, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.3, "rationale": "Cold chain logistics monitoring system using cloud computing and decision trees; applied IT rather than a clear CS category, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7134283222, "categories": ["distributed_computing", "scientific_computing"], "primary": "distributed_computing", "confidence": 0.55, "rationale": "Parallelized Kubernetes/Ray preprocessing and ML workflow for geospatial prediction; distributed processing is the main contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4297517292, "categories": ["cloud_computing", "operating_systems"], "primary": "operating_systems", "confidence": 0.6, "rationale": "Overview of Docker containers and OS-level virtualization and isolation.", "model": "claude-haiku-5-5"}
++{"work_id": 4392175659, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Survey of security threats and countermeasures in cloud computing.", "model": "claude-haiku-5-5"}
++{"work_id": 4401943441, "categories": ["edge_computing", "cloud_computing"], "primary": "edge_computing", "confidence": 0.7, "rationale": "Empirical study of edge-to-cloud continuum deployment, scheduling and load balancing for IoT applications.", "model": "claude-haiku-5-5"}
++{"work_id": 3137445566, "categories": ["sustainable_computing", "computer_architecture"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Comparison of power consumption of rack vs blade servers in a data center, energy-efficiency focus.", "model": "claude-haiku-5-5"}
++{"work_id": 2900680085, "categories": ["cloud_computing", "operating_systems"], "primary": "cloud_computing", "confidence": 0.45, "rationale": "Thematic analysis via LDA of virtualization patents as enabler of cloud computing; bibliometric-style study.", "model": "claude-haiku-5-5"}
++{"work_id": 4403332127, "categories": ["cloud_computing", "sustainable_computing"], "primary": "cloud_computing", "confidence": 0.65, "rationale": "Geographical load balancing across data centers to reduce cloud energy consumption.", "model": "claude-haiku-5-5"}
++{"work_id": 4320024362, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Wasserstein GAN for simulating time-domain gravitational wave signals; generative model training is the core contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7118421194, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.85, "rationale": "Noise optimization to fix mode collapse in text-to-image diffusion models.", "model": "claude-haiku-5-5"}
++{"work_id": 7110667317, "categories": ["llm"], "primary": "llm", "confidence": 0.5, "rationale": "Discusses use of GenAI tools for history teaching; education application of generative AI, not clearly CS research beyond tool use.", "model": "claude-haiku-5-5"}
++{"work_id": 7166546311, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "User behavior study of generative AI's effect on students' skill retention using clustering and predictive modeling of user profiles.", "model": "claude-haiku-5-5"}
++{"work_id": 4379985046, "categories": ["generative_ai", "multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Text-driven facial image editing combining CLIP and StyleGAN2, linking text and images.", "model": "claude-haiku-5-5"}
++{"work_id": 2999632917, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Conditional adversarial network synthesizing PolSAR images across frequency bands.", "model": "claude-haiku-5-5"}
++{"work_id": 4405955397, "categories": ["generative_ai", "computational_biology"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Review of EEG decoding into images, video, audio via generative models; neuroscience signal decoding with generative focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4409570677, "categories": ["ml_theory", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "VAE latent parametrization of CMB power spectra for cosmological models; scientific ML application in physics.", "model": "claude-haiku-5-5"}
++{"work_id": 4413402631, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Qualitative study of students' generative AI use in multicultural classrooms; human-AI interaction and user perceptions.", "model": "claude-haiku-5-5"}
++{"work_id": 4381736219, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.85, "rationale": "Conditional Wasserstein DCGAN proposed for stabilizing GAN training and generating images and mouse trajectories.", "model": "claude-haiku-5-5"}
++{"work_id": 7155650775, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Conceptual education/sociocultural theory paper about generative AI in learning; not CS research in the ontology sense.", "model": "claude-haiku-5-5"}
++{"work_id": 7152054860, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Controlling GAN/VAE image generative models without supervision, latent exploration and scene generation with geometric control.", "model": "claude-haiku-5-5"}
++{"work_id": 4414361486, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Qualitative higher-education study of course coordinators' views on GenAI; social science rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7164539746, "categories": ["generative_ai", "multimodal_ai", "computer_graphics"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Text-and-speech driven 3D talking head generation using SDS text-to-3D and differentiable rendering; spans generation, multimodal input, and graphics.", "model": "claude-haiku-5-5"}
++{"work_id": 7165127855, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Conceptual and survey-based study of AI adoption in history education; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4404404823, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Diffusion bridge model for time series forecasting; diffusion-based generative modeling applied to forecasting.", "model": "claude-haiku-5-5"}
++{"work_id": 7220858792, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Questionnaire-based education study on AIGC teaching design in digital trade programs; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4406954321, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Variational Schr\u00f6dinger momentum diffusion for generative modeling with transport-optimized scores.", "model": "claude-haiku-5-5"}
++{"work_id": 7204970079, "categories": ["llm"], "primary": "llm", "confidence": 0.4, "rationale": "Learning analytics of GenAI virtual patient dialogue logs; primarily educational analytics, with LLM-based dialogue as context.", "model": "claude-haiku-5-5"}
++{"work_id": 4381744251, "categories": ["nlp"], "primary": null, "confidence": 0.2, "rationale": "Unsupervised domain adaptation for image classification with GAN-based features; best fit is weak, likely computer_vision or deep learning, but no category clearly dominates from the abstract. Leaving uncategorized for conservatism.", "model": "claude-haiku-5-5"}
++{"work_id": 4402504098, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Improved CycleGAN for generating metal wear images to expand a sample library; generative model for image synthesis applied to vision data.", "model": "claude-haiku-5-5"}
++{"work_id": 3015491268, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Conditional GAN with mixture density output for probabilistic time series forecasting; generative model design.", "model": "claude-haiku-5-5"}
++{"work_id": 4400444727, "categories": ["generative_ai", "ml_theory"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Evaluation metrics for generative models (FID, IS, likelihood-based protocol), centred on evaluating generative models.", "model": "claude-haiku-5-5"}
++{"work_id": 7127452594, "categories": ["multimodal_ai", "speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Audio forgery localization, detecting forged regions in speech/audio; relates to AI-generated content detection, so generative_ai is also plausible.", "model": "claude-haiku-5-5"}
++{"work_id": 7113405664, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.4, "rationale": "Supplementary materials on human detection of AI-generated images; very little abstract information, so low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 2966462029, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.35, "rationale": "Community consistency tracking in dynamic networks; data-mining/graph analysis, could also fit algorithms or ml_theory.", "model": "claude-haiku-5-5"}
++{"work_id": 7167727964, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Student perspectives and use of GenAI tools in dance education; user study of interactive tool use.", "model": "claude-haiku-5-5"}
++{"work_id": 2810722879, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.65, "rationale": "Extends ambient generative adversarial training to the hidden space of the generator.", "model": "claude-haiku-5-5"}
++{"work_id": 7125473265, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.3, "rationale": "Critical political-economy analysis of recommender systems on platforms; not a clear CS technical contribution, low confidence fairness/governance fit.", "model": "claude-haiku-5-5"}
++{"work_id": 7135177077, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Mixed-methods study of AI-enabled teaching in engineering education with student learning behaviour; human-AI interaction framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4200011910, "categories": ["computer_vision", "generative_ai", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Super-resolution image enhancement via GAN-based ESRGAN variant; image restoration is computer vision, GAN is generative, architecture modification is deep learning architectures.", "model": "claude-haiku-5-5"}
++{"work_id": 7215117915, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Qualitative study of how creators communicate AI use on Instagram; social computing and human-AI interaction, not CS-core method.", "model": "claude-haiku-5-5"}
++{"work_id": 4395661446, "categories": ["computer_vision", "generative_ai"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Ocean wave spectrum retrieval from SAR using a conditional GAN and XGBoost; remote sensing application with GAN-based approach. Primarily earth observation, so low-confidence CS mapping.", "model": "claude-haiku-5-5"}
++{"work_id": 3130253613, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Survey-style discussion of GAN-based image translation application.", "model": "claude-haiku-5-5"}
++{"work_id": 7171857204, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Educational psychology SEM study of teachers' AI literacy and TPACK; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4392008307, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Comparison of generative AI tools for new users; overview of generative AI tools.", "model": "claude-haiku-5-5"}
++{"work_id": 7143415918, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Business management study of generative AI applications based on secondary data; not primarily CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4407424148, "categories": ["generative_ai", "ml_theory", "foundation_models"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Uses conditional diffusion models to generate synthetic data for prediction with statistical guarantees across multimodal tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 7112438470, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Report on GenAI usage in architecture education; educational study rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 7125351784, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Diffusion-based time series generation with a Mixture-of-Experts neural CDE architecture.", "model": "claude-haiku-5-5"}
++{"work_id": 4406930504, "categories": ["generative_ai", "scientific_computing", "computational_biology"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Guidance for diffusion-based molecular geometry generation applied to drug design; generative model is the core contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4412083660, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.7, "rationale": "AI image generation of cultural festivals plus a public perception survey on the uncanny valley.", "model": "claude-haiku-5-5"}
++{"work_id": 4387666097, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Color transfer between images via optimal transport is image synthesis/manipulation; could also be computer vision.", "model": "claude-haiku-5-5"}
++{"work_id": 7118194253, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Education study on AI literacy of adult learners; user-focused evaluation rather than a computing contribution per se.", "model": "claude-haiku-5-5"}
++{"work_id": 7127651506, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Uses AIGC to create visual cultural content for new media; more cultural/design study than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4367292510, "categories": ["generative_ai", "computer_graphics"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Deep convolutional GAN for generating ship hull geometries as a parametric modeller.", "model": "claude-haiku-5-5"}
++{"work_id": 2904629716, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Attentional GAN for large-scale single image super-resolution, an image restoration task.", "model": "claude-haiku-5-5"}
++{"work_id": 4406160051, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.4, "rationale": "Magazine piece on generative AI for scientific image reconstruction; little detail in abstract.", "model": "claude-haiku-5-5"}
++{"work_id": 7138839474, "categories": ["computer_vision", "generative_ai"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Sketch recognition classifier using attention and CycleGAN features; primarily a recognition task.", "model": "claude-haiku-5-5"}
++{"work_id": 7133211554, "categories": ["ml_theory", "algorithms"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Approximation of optimal transport distances via projections; mathematical methodology for distribution comparison.", "model": "claude-haiku-5-5"}
++{"work_id": 7116068851, "categories": ["ai_safety", "llm", "ml_theory"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Interpretability assistants that detect jailbreaks and latent attributes in LLM activations; primarily interpretability/safety of LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4417167881, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Management/governance study of AI adoption and accountability; not CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 2910154596, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Ethics analysis of a vehicle product against a software engineering code; ethics essay rather than CS technical research, though it touches software engineering ethics.", "model": "claude-haiku-5-5"}
++{"work_id": 7154958542, "categories": ["nlp"], "primary": "nlp", "confidence": 0.45, "rationale": "BERT-based review text mining of AI companion apps, mainly consumer behaviour framing.", "model": "claude-haiku-5-5"}
++{"work_id": 7133246977, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Survey study of student attitudes toward AI in education; social science rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7119528090, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Legal analysis of criminal liability for GenAI providers; legal scholarship.", "model": "claude-haiku-5-5"}
++{"work_id": 7168242482, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Landscape analysis of ethics in computing education; education/policy focus, not a CS technical category.", "model": "claude-haiku-5-5"}
++{"work_id": 7151628917, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Governance framework for deceptive AI research risk levels.", "model": "claude-haiku-5-5"}
++{"work_id": 4399042091, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Ethical analysis of health recommender systems for older adults; closest to human-centred interactive systems, but largely ethics.", "model": "claude-haiku-5-5"}
++{"work_id": 7214017973, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Academic design case on responsible generative AI use in doctoral education; education and AI ethics rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 2803572453, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Studies verbatim content republishing networks across news producers; closer to information/news ecosystem analysis with text analysis, so tentative assignment.", "model": "claude-haiku-5-5"}
++{"work_id": 3118441209, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Ethics of web analytics and user behaviour monitoring in online retail; closest to user-facing interactive systems and privacy-related ethics, not clearly CS core.", "model": "claude-haiku-5-5"}
++{"work_id": 4396786581, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Shapley-value feature importance and fairness detection in ML models, a principled methodology for interpretability and fairness.", "model": "claude-haiku-5-5"}
++{"work_id": 7161719656, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.55, "rationale": "Discusses AI agents generating hypotheses and papers and the epistemology of machine-driven science.", "model": "claude-haiku-5-5"}
++{"work_id": 7164757242, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Conceptual education policy analysis of algorithmic recommendation governance; social science rather than computer science core.", "model": "claude-haiku-5-5"}
++{"work_id": 7119498334, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Policy and governance discourse analysis of AI regulation across jurisdictions; not a CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7221224735, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Community service workshop on using NotebookLM in Islamic religious education; educational practice rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4400082893, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Survey study of students' security and ethical attitudes toward AI in education; educational psychology rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 4394613885, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Document analysis for educational policy development on AI community initiative; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4415318303, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Benchmark evaluating LLM psychogenicity, delusion reinforcement and harm enablement, a safety evaluation of LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4386791501, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Ethical review and risk mitigation of AI research involving human subjects; governance-oriented, closest to AI governance under ai_safety.", "model": "claude-haiku-5-5"}
++{"work_id": 4411990784, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Questionnaire survey of student AI awareness; social science rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4220714491, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Curriculum analysis of informatics education; not a CS technical contribution in the ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 3096804093, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Questionnaire study on AI implementation in African governments; policy/e-government rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 4405384079, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Interview study of communication researchers' wellbeing; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3093512442, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Editorial guidance for authors of healthcare AI manuscripts; publishing practice rather than a CS research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7147573305, "categories": ["llm", "ai_safety", "ml_theory"], "primary": "llm", "confidence": 0.7, "rationale": "Membership inference on LLMs for memorization in financial forecasting; LLM evaluation with memorization/contamination auditing as safety-related diagnostics.", "model": "claude-haiku-5-5"}
++{"work_id": 4415829954, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Survey of academics' AI competency in Nigeria; social/educational study rather than CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4392109933, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Management discussion on AI in research process; not a CS technical contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7211974235, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Essay on AI and human communication in education and society; communication studies rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 3110765181, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Social impacts, harms, bias and governance of AI; not a CS technical contribution, but closest category is AI safety/fairness/governance.", "model": "claude-haiku-5-5"}
++{"work_id": 4399365038, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Algorithm audit of ChatGPT content moderation guardrails and biases, evaluating LLM outputs for harmful content.", "model": "claude-haiku-5-5"}
++{"work_id": 4386555180, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.75, "rationale": "Discusses limitations of XAI within responsible AI covering privacy, fairness, contestability.", "model": "claude-haiku-5-5"}
++{"work_id": 4283010037, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Network science and social inequality chapter; primarily social/complex systems rather than a clear CS category, though fairness of ranking algorithms is touched on.", "model": "claude-haiku-5-5"}
++{"work_id": 4412611638, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Conceptual paper on AGI architecture and analogical reasoning; loosely about general intelligence systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4290803082, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Legal and policy analysis of data sovereignty and colonialism; not clearly computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7204733568, "categories": [], "primary": null, "confidence": 0.35, "rationale": "Education policy survey of stakeholder preferences on AI in schools; social science rather than CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7162150262, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.85, "rationale": "Evaluates LLM agent behaviour under authority pressure for safety, using open-source LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 4390091701, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Argues for moral dimensions beyond fairness in assessing and designing AI systems; AI ethics and fairness.", "model": "claude-haiku-5-5"}
++{"work_id": 4399331704, "categories": [], "primary": null, "confidence": 0.35, "rationale": "Decision-making study using balanced scorecard and fuzzy sets for AI in finance; management science rather than a CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7134710523, "categories": ["ai_safety", "hci"], "primary": "hci", "confidence": 0.5, "rationale": "Conceptual mental model of explanations for users of AI systems; explainability and user perception framing fits HCI with an AI-transparency angle.", "model": "claude-haiku-5-5"}
++{"work_id": 4399198750, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Measuring and mitigating discrimination/bias in tabular datasets with multiple protected attributes is fairness work.", "model": "claude-haiku-5-5"}
++{"work_id": 7135082306, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Mechanistic interpretability of safety behavior in Llama-3-70B; LLM-centred and safety-focused.", "model": "claude-haiku-5-5"}
++{"work_id": 4405576526, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Survey of AI regulation and governance landscape in New Zealand.", "model": "claude-haiku-5-5"}
++{"work_id": 4411795627, "categories": ["ai_safety", "hci"], "primary": "ai_safety", "confidence": 0.4, "rationale": "Critical theoretical reflection on algorithmic bias and justice in inclusive math education; education-focused with fairness and accessibility themes.", "model": "claude-haiku-5-5"}
++{"work_id": 7164020700, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.3, "rationale": "Discusses AI-assisted aesthetic production and originality/copyright concerns; weak, popular-style piece on generative AI in art.", "model": "claude-haiku-5-5"}
++{"work_id": 7134826009, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.35, "rationale": "AI governance framework for higher education institutions; governance focus but primarily education policy study.", "model": "claude-haiku-5-5"}
++{"work_id": 7165690142, "categories": ["ai_safety", "hci"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Red teaming of ChatGPT with faith leaders for safety and accountability; LLM safety evaluation with participatory HCI elements.", "model": "claude-haiku-5-5"}
++{"work_id": 7117162136, "categories": ["ai_safety", "llm", "ai_agents"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Monitoring chain-of-thought of reasoning models and agents for misbehavior; safety evaluation of frontier LLMs.", "model": "claude-haiku-5-5"}
++{"work_id": 2997003061, "categories": ["ai_safety"], "primary": null, "confidence": 0.2, "rationale": "Model-based safety design for civil aircraft systems using STPA; safety engineering rather than CS-core AI, best fit is software engineering or none, left without a firm category.", "model": "claude-haiku-5-5"}
++{"work_id": 7133303346, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Fully digital MPPT circuit for piezoelectric energy harvesting interface; hardware/circuit design, energy harvesting, closest to low-power hardware design.", "model": "claude-haiku-5-5"}
++{"work_id": 2950925562, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Pure mathematics (geometric measure theory, Minkowski content); not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3115601909, "categories": ["sustainable_computing"], "primary": null, "confidence": 0.25, "rationale": "Peer-to-peer energy trading auction in smart grid; mainly power systems and market mechanism design, weakly related to CS algorithmic game theory; low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4384570508, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.35, "rationale": "Scalable data archiving system design with data placement and metadata management; vague, closest to storage/data management.", "model": "claude-haiku-5-5"}
++{"work_id": 3187190559, "categories": ["distributed_computing", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Multi-objective workflow scheduling in cloud using neural network and evolutionary algorithm; cloud resource scheduling.", "model": "claude-haiku-5-5"}
++{"work_id": 7128410034, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Computer algebra algorithms for parametric polynomial system solving via rational univariate representation; symbolic computation.", "model": "claude-haiku-5-5"}
++{"work_id": 3006062179, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Simulation of ML system performance via offline profiling and parallelization strategies; ML systems performance modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 2860211826, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Fractional-order fault-tolerant control of an isotope separation process; control engineering rather than a CS category, weakly cyber-physical.", "model": "claude-haiku-5-5"}
++{"work_id": 7133367970, "categories": ["cybersecurity", "distributed_computing"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Deep learning for failure prediction from HPC log data; fault tolerance in HPC systems.", "model": "claude-haiku-5-5"}
++{"work_id": 2907696829, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Energy management of microgrid with hydrogen and MPC; power and energy systems engineering, weakly sustainable computing.", "model": "claude-haiku-5-5"}
++{"work_id": 2918759589, "categories": ["distributed_computing", "computer_architecture"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Special issue on high-performance and distributed computing covering parallel applications, GPU computing, energy and cloud computing.", "model": "claude-haiku-5-5"}
++{"work_id": 2916417520, "categories": ["distributed_computing", "programming_languages"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Programming model and runtime for heterogeneous multi-GPU/CPU HPC systems and data locality.", "model": "claude-haiku-5-5"}
++{"work_id": 4413272460, "categories": ["blockchain", "distributed_computing"], "primary": "blockchain", "confidence": 0.85, "rationale": "Blockchain consensus, sharding and decentralized storage framework.", "model": "claude-haiku-5-5"}
++{"work_id": 2805585446, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Engineering education study using cluster analysis of student performance; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4409660083, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Inverter model-predictive control in power electronics; control engineering rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7161207400, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Short HAL news item about instrumentation equipment; not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 4400690093, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.9, "rationale": "Survey of Layer 2 rollup scaling solutions for blockchains.", "model": "claude-haiku-5-5"}
++{"work_id": 4388580994, "categories": ["computer_architecture", "scientific_computing"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Performance evaluation of a CPU architecture (Xeon MAX) for bandwidth-bound HPC workloads.", "model": "claude-haiku-5-5"}
++{"work_id": 2904849773, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Workflow-based business simulation training platform for education; weakly fits applied systems, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4401163816, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Frequency estimation signal processing for proton magnetometer instrumentation; measurement engineering rather than CS.", "model": "claude-haiku-5-5"}
++{"work_id": 7201761843, "categories": ["distributed_computing", "database_systems"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Multi-cluster Kafka geo-replication and failover library; distributed storage/middleware and data replication across clusters.", "model": "claude-haiku-5-5"}
++{"work_id": 3152708525, "categories": ["distributed_computing", "computer_networks"], "primary": "distributed_computing", "confidence": 0.7, "rationale": "Server selection and optimistic synchronization for distributed processing to minimize end-to-end delay.", "model": "claude-haiku-5-5"}
++{"work_id": 4200525533, "categories": ["nlp", "ai_safety"], "primary": "nlp", "confidence": 0.55, "rationale": "Deep learning text classification of social media messages for suicide risk; mental health application, primarily text analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 3038439485, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational R&D on pop-up book media for early reading; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2972427306, "categories": ["distributed_computing", "deep_learning_architectures"], "primary": "distributed_computing", "confidence": 0.65, "rationale": "Accelerating MapReduce and Spark iterative ML workloads with GPUs via adaptive scheduling; distributed/parallel systems with hardware acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 3014911997, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Finite-time synchronization theory for fractional-order memristor neural networks; control-theoretic analysis of neural network dynamics.", "model": "claude-haiku-5-5"}
++{"work_id": 2998437218, "categories": ["distributed_computing", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Ontology-based resource monitoring and fault diagnosis for cloud systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4416926542, "categories": ["cloud_computing", "sustainable_computing", "distributed_computing"], "primary": "sustainable_computing", "confidence": 0.75, "rationale": "Energy-efficient job scheduling in green cloud data centers using metaheuristics.", "model": "claude-haiku-5-5"}
++{"work_id": 2903036202, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Control design and LMI-based synchronization of discrete-time Lur'e systems; systems and control theory closest to ml_theory-free categories, low fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4286377584, "categories": ["ml_theory", "scientific_computing"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Empirical Bayesian estimators for SAR coherence; statistical estimation methodology applied to remote sensing.", "model": "claude-haiku-5-5"}
++{"work_id": 7131086620, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Survey of radiologists about AI adoption in breast imaging; medical/health-services study rather than a CS contribution, so no ontology category clearly applies.", "model": "claude-haiku-5-5"}
++{"work_id": 3045355300, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Distributed adaptive consensus control for multiagent systems; control-theoretic rather than a clear fit to listed categories, closest to ml_theory-style principled adaptation but arguably outside the ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 4405718414, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Revisits dynamic software change management for safely updating distributed systems, a software evolution topic.", "model": "claude-haiku-5-5"}
++{"work_id": 2903731864, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Statistical simulation study of count data models; statistics rather than a computer-science ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 4200552233, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Performance monitoring tooling for Python applications; closest to performance analysis and tooling in software engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 7216249722, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Indonesian literature education research on e-modules; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2891039752, "categories": ["distributed_computing", "cloud_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Hypervisor-level checkpoint-restart for MPI clusters of VMs provides fault tolerance for distributed applications in cloud environments.", "model": "claude-haiku-5-5"}
++{"work_id": 4403092717, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Parallel processing of geographical data on mobile devices using parallel streams; parallel computing focus.", "model": "claude-haiku-5-5"}
++{"work_id": 4407784105, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Parallel subgraph-level distributed training of GNNs on heterogeneous platforms; distributed ML training systems, with GNN architecture as context.", "model": "claude-haiku-5-5"}
++{"work_id": 4411552467, "categories": ["distributed_computing", "blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Improved PBFT consensus mechanism for a medical device traceability blockchain; blockchain consensus system.", "model": "claude-haiku-5-5"}
++{"work_id": 2887272015, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Dynamic malleability of HPC jobs and resource management for parallel MPI/OpenMP applications; fits parallel/HPC systems software.", "model": "claude-haiku-5-5"}
++{"work_id": 3123252765, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Servo motion control and vibration suppression for motion systems; closest to robotics/motion control, though it is control engineering.", "model": "claude-haiku-5-5"}
++{"work_id": 3093694272, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Information systems management research on IT affordances and organisational innovation; not computer science category.", "model": "claude-haiku-5-5"}
++{"work_id": 4411584530, "categories": ["reinforcement_learning", "distributed_computing"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Distributional RL algorithm applied to pod scheduling in satellite clusters; cloud/cluster resource scheduling also applies.", "model": "claude-haiku-5-5"}
++{"work_id": 4383532084, "categories": ["database_systems", "distributed_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Disk-based erasure-coded archival storage system for LHC data; distributed storage middleware.", "model": "claude-haiku-5-5"}
++{"work_id": 3048509415, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Literature review on manufacturing quality improvement philosophies; industrial engineering rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4414878360, "categories": ["programming_languages", "distributed_computing"], "primary": "programming_languages", "confidence": 0.8, "rationale": "Domain-specific language for mapping distributed programs; DSL design with parallel runtime mapping.", "model": "claude-haiku-5-5"}
++{"work_id": 4410853603, "categories": ["information_retrieval", "distributed_computing", "nlp"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "AI-driven system for HPC resource access and scheduling with a fine-tuned RoBERTa chatbot; mixed focus, primarily HPC resource management.", "model": "claude-haiku-5-5"}
++{"work_id": 4407673169, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Self-healing fault-tolerant framework for cloud distributed systems and platform reliability.", "model": "claude-haiku-5-5"}
++{"work_id": 3011113897, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Multi-stage distributed big-data processing framework evaluating agent connection topologies.", "model": "claude-haiku-5-5"}
++{"work_id": 3128655982, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Improved DDPG for obstacle avoidance of a single agent; RL algorithm contribution applied to robot-like navigation.", "model": "claude-haiku-5-5"}
++{"work_id": 2996458828, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Meta-RL with subtask graph inference and intrinsic exploration reward.", "model": "claude-haiku-5-5"}
++{"work_id": 4403461025, "categories": ["reinforcement_learning", "scientific_computing"], "primary": "reinforcement_learning", "confidence": 0.65, "rationale": "Adaptive dynamic programming (RL-based) controller for wind turbine MPPT; control application.", "model": "claude-haiku-5-5"}
++{"work_id": 7160928702, "categories": ["ai_agents", "llm", "reinforcement_learning"], "primary": "ai_agents", "confidence": 0.8, "rationale": "LLM agents with external skill management in agentic RL.", "model": "claude-haiku-5-5"}
++{"work_id": 7162259189, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Evaluates task complexity metrics for RL on robotic manipulation tasks.", "model": "claude-haiku-5-5"}
++{"work_id": 4388561956, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Comparison and extension of hierarchical RL (MAXQ) versus Q-learning on a game.", "model": "claude-haiku-5-5"}
++{"work_id": 4285172767, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Frame-skipping DRL agent performance improvement using hidden observations.", "model": "claude-haiku-5-5"}
++{"work_id": 4400838813, "categories": ["reinforcement_learning", "nlp"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "DDPG-based portfolio management; finance application of deep RL with GCN/TCN features.", "model": "claude-haiku-5-5"}
++{"work_id": 4416874462, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Off-policy average-reward actor-critic algorithm (SAC variant).", "model": "claude-haiku-5-5"}
++{"work_id": 7138837252, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.75, "rationale": "Reward-free RL theory and practice, career highlight summarizing RL research.", "model": "claude-haiku-5-5"}
++{"work_id": 4412985292, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "PPO-based variant for robot path planning in grid mazes with height-based reward shaping; RL algorithm contribution applied to robot path planning.", "model": "claude-haiku-5-5"}
++{"work_id": 4384135373, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Explainable RL via fuzzy neural networks and generative model on Atari; explainability of RL agents.", "model": "claude-haiku-5-5"}
++{"work_id": 4413402209, "categories": ["robotics", "reinforcement_learning", "algorithms"], "primary": "robotics", "confidence": 0.6, "rationale": "UAV quadrotor path planning combining MDP with improved artificial potential field; path planning for a robotic system.", "model": "claude-haiku-5-5"}
++{"work_id": 4411020042, "categories": ["algorithms", "reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Contextual combinatorial bandits with regret bounds; online learning theory for bandits.", "model": "claude-haiku-5-5"}
++{"work_id": 4414230988, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "SAC algorithm enhanced with PER and entropy tuning for autonomous vehicle collision avoidance in CARLA.", "model": "claude-haiku-5-5"}
++{"work_id": 3177555056, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Dissertation on general value functions, successor representations and temporal-difference learning for predictive RL.", "model": "claude-haiku-5-5"}
++{"work_id": 4400812532, "categories": ["robotics", "reinforcement_learning"], "primary": "robotics", "confidence": 0.7, "rationale": "Mobile robot control with ROS, SLAM and actor-critic control, tested on real robots.", "model": "claude-haiku-5-5"}
++{"work_id": 4405571822, "categories": ["robotics", "reinforcement_learning"], "primary": "robotics", "confidence": 0.7, "rationale": "Comparison of DRL algorithms for tilt-rotor UAV hover-to-cruise transition; aerial robotics application.", "model": "claude-haiku-5-5"}
++{"work_id": 7214803104, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Multi-objective RL methods for continuous control, including adaptive scalarisation and replay mechanisms.", "model": "claude-haiku-5-5"}
++{"work_id": 4392858396, "categories": ["nlp", "hci"], "primary": "hci", "confidence": 0.4, "rationale": "Game-based AI education for rural students; educational and user-facing study rather than a core AI method.", "model": "claude-haiku-5-5"}
++{"work_id": 3040890460, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Contextual bandit variant with missing rewards addresses bandit learning, a core RL subfield; clustering-based estimation is methodological.", "model": "claude-haiku-5-5"}
++{"work_id": 4405358450, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Review of open educational resources for music history teaching; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4388579628, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Soft actor-critic applied to automatic parking path planning and tracking; RL method applied to vehicle robotics.", "model": "claude-haiku-5-5"}
++{"work_id": 7160554706, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Review of dopamine reward prediction error in neuroscience; history of RL influence but primarily neuroscience, not CS research.", "model": "claude-haiku-5-5"}
++{"work_id": 3216293753, "categories": ["ml_theory", "ai_agents"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Active inference and tree-search planning framed as structure learning; theoretical planning framework, borderline with agents.", "model": "claude-haiku-5-5"}
++{"work_id": 4313459177, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Evolutionary deep RL with soft actor-critic for Volt-VAR control in power distribution networks; RL applied to power systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3014616672, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Monte Carlo Q-value function for multi-agent decision making under Dec-POMDP; multi-agent planning and value methods.", "model": "claude-haiku-5-5"}
++{"work_id": 7213255109, "categories": ["reinforcement_learning", "scientific_computing"], "primary": "reinforcement_learning", "confidence": 0.4, "rationale": "MDP abstraction algorithm for policy interpretability in conservation decision-making; MDP-based planning with ecological application.", "model": "claude-haiku-5-5"}
++{"work_id": 7151096044, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.9, "rationale": "Multi-agent actor-critic (MAPPO) stabilisation via self-distillation; core MARL algorithm contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 7167869540, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Dynamic programming for decentralized POMDPs with delayed information sharing; multi-agent decision theory.", "model": "claude-haiku-5-5"}
++{"work_id": 4408611127, "categories": ["reinforcement_learning", "computer_vision"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Studies DQN state representations for image-based RL agents in Snake; core is RL with image observations.", "model": "claude-haiku-5-5"}
++{"work_id": 3208966735, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Intrinsic-motivation exploration for cooperative multi-agent RL under sparse rewards.", "model": "claude-haiku-5-5"}
++{"work_id": 4385764347, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.75, "rationale": "State abstraction for RL with information-theoretic structural entropy framework.", "model": "claude-haiku-5-5"}
++{"work_id": 4412795647, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Multi-agent deep RL (QMIX variant) for truck platoon coordination; applied transportation domain.", "model": "claude-haiku-5-5"}
++{"work_id": 7123339254, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Constrained average-reward MDPs with partial observability; LP duality analysis of MDPs.", "model": "claude-haiku-5-5"}
++{"work_id": 7221661943, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Actor-critic RL for distributed cooperative output regulation of multi-agent control systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4391768605, "categories": ["llm", "reinforcement_learning", "hci"], "primary": "llm", "confidence": 0.7, "rationale": "User study of ChatGPT assisting users in designing RL state/reward for traffic control.", "model": "claude-haiku-5-5"}
++{"work_id": 3170511602, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Cooperative exploration method for multi-agent deep RL.", "model": "claude-haiku-5-5"}
++{"work_id": 4414610855, "categories": ["reinforcement_learning", "nlp"], "primary": "reinforcement_learning", "confidence": 0.5, "rationale": "Q-learning based personalized recommendation for English vocabulary teaching; educational application.", "model": "claude-haiku-5-5"}
++{"work_id": 4406073402, "categories": ["reinforcement_learning", "ai_safety"], "primary": "reinforcement_learning", "confidence": 0.75, "rationale": "Concept-bottleneck interpretable RL policies with label-efficient training; interpretability framed for RL.", "model": "claude-haiku-5-5"}
++{"work_id": 3174702694, "categories": ["computer_vision", "reinforcement_learning", "robotics"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Embodied active learning for semantic segmentation using deep RL in a 3D simulator; spans vision, RL and robotic embodiment.", "model": "claude-haiku-5-5"}
++{"work_id": 4402437985, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Two-tier multi-agent RL learning strategy and tactics in competitive games.", "model": "claude-haiku-5-5"}
++{"work_id": 2793827368, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.65, "rationale": "Adaptive dynamic programming with critic networks for optimal control of uncertain nonlinear systems; control-theoretic RL with stability guarantees.", "model": "claude-haiku-5-5"}
++{"work_id": 4400526476, "categories": ["reinforcement_learning", "robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "Offline-to-online RL applied to quadrotor docking control.", "model": "claude-haiku-5-5"}
++{"work_id": 7213353131, "categories": ["ai_agents", "autonomous_scientific_discovery"], "primary": "ai_agents", "confidence": 0.55, "rationale": "Recursive self-improvement of autonomous agent exploration via simulated discovery trees.", "model": "claude-haiku-5-5"}
++{"work_id": 3217267205, "categories": ["reinforcement_learning", "information_retrieval"], "primary": "reinforcement_learning", "confidence": 0.4, "rationale": "Probabilistic simulator of spatial retail demand used to train Deep Q-learning allocation policy; retail/economics framing is partly outside CS categories.", "model": "claude-haiku-5-5"}
++{"work_id": 7127355566, "categories": ["reinforcement_learning", "deep_learning_architectures"], "primary": "reinforcement_learning", "confidence": 0.75, "rationale": "Improves DQN target update mechanism with convergence analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 7117961497, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Actor-critic RL with Lyapunov certificates for stabilizing control.", "model": "claude-haiku-5-5"}
++{"work_id": 4393861500, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Stochastic control and free boundary PDE analysis of investment; mathematics/economics rather than CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 6884998335, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Convergence and regret analysis of q-learning for controlled diffusions.", "model": "claude-haiku-5-5"}
++{"work_id": 4220992898, "categories": ["computer_architecture", "deep_learning_architectures", "robotics"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Bio-inspired haptic-temperature spiking neuron built from piezoresistive sensor and VO2 memristor; hardware device for sensory neurons, with e-skin and neuro-robotics framing.", "model": "claude-haiku-5-5"}
++{"work_id": 3011341896, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "CNN compression and sparsity analysis on a hardware CNN accelerator model with processing time estimation.", "model": "claude-haiku-5-5"}
++{"work_id": 3012645150, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Photonic CNN architecture implemented with integrated optics for neural network hardware acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 4388847637, "categories": ["ai_accelerators", "computer_networks"], "primary": "ai_accelerators", "confidence": 0.45, "rationale": "Survey of optical neural networks with discussion of optical communication and 5G applications; hardware-oriented neural computing.", "model": "claude-haiku-5-5"}
++{"work_id": 3047966047, "categories": ["computer_architecture", "ai_accelerators"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Defect characterization of spintronic neuromorphic circuits, a hardware testing and reliability study.", "model": "claude-haiku-5-5"}
++{"work_id": 4416096548, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Curriculum reform for AI and law education; not primarily a computer-science research contribution in a technical category.", "model": "claude-haiku-5-5"}
++{"work_id": 4414291128, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Monolithically integrated on-chip optical neural network for end-to-end inference with fabricated hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4400641184, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Spin-torque oscillator reservoir computing system for energy-efficient neuromorphic hardware time series prediction.", "model": "claude-haiku-5-5"}
++{"work_id": 7133525837, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Mathematical analysis of approximate synchronization in memristive Hopfield networks with proofs of bounds and convergence.", "model": "claude-haiku-5-5"}
++{"work_id": 2989832056, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.55, "rationale": "Semiconductor laser reservoir computing with computational capacity analysis for photonic neuromorphic hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 7138209559, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Memristive neural circuit hardware design for neuromorphic implementation with circuit analysis and PCB validation; hardware-oriented, closest to computer_architecture/neuromorphic hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4312821358, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.7, "rationale": "Analysis of 1S1R crossbar arrays for in-memory VMM operations and wire resistance effects; hardware design of memory arrays.", "model": "claude-haiku-5-5"}
++{"work_id": 4387594915, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.4, "rationale": "Photonic neural network hardware for efficient neural computation; title only, low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7165771277, "categories": ["ai_accelerators", "deep_learning_architectures", "speech_audio"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Spiking transformer architecture for automatic modulation recognition on neuromorphic chip; primarily an SNN architecture applied to radio signals.", "model": "claude-haiku-5-5"}
++{"work_id": 4392976059, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Book/product review in librarianship; not computer-science research.", "model": "claude-haiku-5-5"}
++{"work_id": 3193775649, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.45, "rationale": "Memristive synapse hardware interface for neural signal processing; hardware-software system, neuroprosthetics context.", "model": "claude-haiku-5-5"}
++{"work_id": 2912636021, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Low-power QCA circuit design for a MAC unit used in DSP; digital hardware design.", "model": "claude-haiku-5-5"}
++{"work_id": 4408280919, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Memristor-coupled chaotic neural networks with hardware implementation and image encryption; mainly nonlinear dynamics and circuits.", "model": "claude-haiku-5-5"}
++{"work_id": 3039564587, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Reliability and switching characterization of h-BN memristor devices; device-level hardware study.", "model": "claude-haiku-5-5"}
++{"work_id": 2955631546, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.45, "rationale": "Device model of current-controlled negative differential resistance for neuromorphic building blocks; device/circuit-level hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 2846788133, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Chaotic oscillator circuits with memristor-like diode bridges; analog circuit design fits computer_architecture/hardware, though it is closer to electronics than CS core topics.", "model": "claude-haiku-5-5"}
++{"work_id": 3191516184, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Tutorial on programmable photonic processors, a general-purpose hardware platform for computing, closest to hardware architecture and accelerators.", "model": "claude-haiku-5-5"}
++{"work_id": 2902787482, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Memristive current mirror circuit design for analog and neuromorphic circuits, a hardware circuit-level contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 3091740059, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Memristive crossbar threshold logic circuit design with CMOS comparators; digital logic hardware design.", "model": "claude-haiku-5-5"}
++{"work_id": 7084748818, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Energy-efficient neuromorphic hardware core for self-attention, a hardware accelerator for transformer-style models.", "model": "claude-haiku-5-5"}
++{"work_id": 4405521075, "categories": ["computational_biology", "computer_architecture"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Digital FPGA model of calcium dynamics in a synaptic bouton; biological modelling implemented in hardware, primarily neuroscience modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 4206136645, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "VHDL design space exploration framework for memristive crossbar arrays, an EDA/hardware design tool.", "model": "claude-haiku-5-5"}
++{"work_id": 4225485808, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "All-optical activation function for optical neural network hardware, evaluated on MNIST; hardware for neural computation.", "model": "claude-haiku-5-5"}
++{"work_id": 4410384882, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Integrated all-optical diffractive deep neural network chip, optical hardware for neural network computation.", "model": "claude-haiku-5-5"}
++{"work_id": 4391086735, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Silicon photonic coherent crossbar for universal linear optics, a photonic hardware architecture with no explicit ML framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4323823183, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Black-box simulation of optical neural network systems, hardware-oriented photonic computing; not clearly a fit for a single category but leans to optical AI hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4391451283, "categories": ["ai_accelerators", "edge_computing", "computational_biology"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Diffractive photonic computing units for EEG classification; hardware accelerator for edge seizure detection, with a biomedical application.", "model": "claude-haiku-5-5"}
++{"work_id": 3047123852, "categories": ["computer_architecture", "formal_verification"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Behavioral VHDL modelling of memristive crossbar arrays for in-memory computing; hardware design and verification.", "model": "claude-haiku-5-5"}
++{"work_id": 3143061595, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Overview of AI chip development toward AGI chips, edge AI implementations and design flow.", "model": "claude-haiku-5-5"}
++{"work_id": 3153206747, "categories": ["ai_accelerators", "computer_networks"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Optical microcomb-based convolutional accelerator for neural networks, plus high-rate optical data transmission.", "model": "claude-haiku-5-5"}
++{"work_id": 4408740885, "categories": ["ai_accelerators", "ml_theory"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Photonic hardware for Bayesian neural network probabilistic computation and uncertainty quantification.", "model": "claude-haiku-5-5"}
++{"work_id": 4413066757, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Optical network-on-chip interconnect for memristor arrays in neuromorphic chips.", "model": "claude-haiku-5-5"}
++{"work_id": 7134888657, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Programmable large-scale optical neural network with reservoir computing for image classification.", "model": "claude-haiku-5-5"}
++{"work_id": 4240155423, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.65, "rationale": "Photonic convolutional accelerator hardware for deep optical neural networks.", "model": "claude-haiku-5-5"}
++{"work_id": 4413807734, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "All-optical synapse device for neuromorphic computing hardware.", "model": "claude-haiku-5-5"}
++{"work_id": 4412921988, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Review of photonic reservoir computing, a neuromorphic/optical hardware approach to brain-inspired computing for AI.", "model": "claude-haiku-5-5"}
++{"work_id": 7164587856, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Photonic chip implementation of a spiking graph neural network architecture for energy-efficient AI acceleration.", "model": "claude-haiku-5-5"}
++{"work_id": 6929467513, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Formal analysis of a generalized threshold logic neuron model with input-dependent weights and classification properties.", "model": "claude-haiku-5-5"}
++{"work_id": 4400662090, "categories": ["ai_accelerators", "scientific_computing"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Photonic integrated processor for matrix inversion with I/O efficiency, applied to reservoir training and MIMO precoding, a hardware accelerator contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4366823742, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Photonic tensor core for neural network vector-matrix multiplication with a fabricated chip.", "model": "claude-haiku-5-5"}
++{"work_id": 7171493874, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.35, "rationale": "Photonic integrated circuit with self-configuring networks processing partially coherent light; hardware-oriented with weak AI framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4386996793, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Geophysical InSAR study of fault locking depth; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 2801359411, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Computational neuroscience modeling of synaptic plasticity with HMMs and information measures.", "model": "claude-haiku-5-5"}
++{"work_id": 2946793717, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Circuit-level modeling of memristor devices and parasitic effects; hardware device analysis.", "model": "claude-haiku-5-5"}
++{"work_id": 3112050902, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Control-theoretic synchronization analysis of memristive neural networks with event-triggered impulsive control and Lyapunov conditions.", "model": "claude-haiku-5-5"}
++{"work_id": 4383046458, "categories": ["ai_safety", "reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.5, "rationale": "Data-driven NMPC with learned dynamics and safety barrier functions; control-oriented, with learning of control policy, so mainly control/learning; RL not central but learning-based control fits reinforcement_learning loosely. Low-confidence multi-label.", "model": "claude-haiku-5-5"}
++{"work_id": 2804990859, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.9, "rationale": "#P-completeness of counting homomorphisms from fundamental groups; complexity hardness result.", "model": "claude-haiku-5-5"}
++{"work_id": 2795888755, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.3, "rationale": "Algebraic study of finite generation and presentations of monoid acts; pure mathematics rather than CS, so at most weakly related to algorithms/theory.", "model": "claude-haiku-5-5"}
++{"work_id": 7205015363, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.35, "rationale": "Automata theory applied to DNA splicing limit languages; formal language theory, closest to automata theory under computational_complexity.", "model": "claude-haiku-5-5"}
++{"work_id": 2979317633, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Biomedical spectroscopy study of skin AGE measurement; not computer science research primarily.", "model": "claude-haiku-5-5"}
++{"work_id": 3037346090, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Computationally efficient inverse-free Kalman filter via approximate matrix inverse; numerical/estimation algorithm, closest to algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 2947814064, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Movable cellular automata simulating biological systems; biological simulation modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 4312278247, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.25, "rationale": "Signal processing measure of time variance for filter banks; weak fit to any category, closest is signal/filter-bank architecture theory.", "model": "claude-haiku-5-5"}
++{"work_id": 2928354599, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.55, "rationale": "Co-evolving cellular automata modelling morphogenesis and growth factors; biological modelling.", "model": "claude-haiku-5-5"}
++{"work_id": 2791221456, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Fuzzy Tsukamoto prediction system for fire-prone areas; applied fuzzy-logic prediction, weakly fitting a probabilistic/predictive modelling category.", "model": "claude-haiku-5-5"}
++{"work_id": 4410227679, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.5, "rationale": "IMM maneuvering target tracking with deep learning priors; tracking is a sensing/estimation task typically in robotics/autonomous systems, with a learning component.", "model": "claude-haiku-5-5"}
++{"work_id": 2936835067, "categories": ["computational_complexity", "applied_cryptography"], "primary": "computational_complexity", "confidence": 0.8, "rationale": "Unique Games hardness of approximating cumulative pebbling cost; a hardness result for a computational problem motivated by memory-hard functions.", "model": "claude-haiku-5-5"}
++{"work_id": 4323338639, "categories": ["algorithms", "deep_learning_architectures"], "primary": "algorithms", "confidence": 0.6, "rationale": "ADMM-based scalable algorithm for regularized minimax approximation, applied to filter design and randomized neural networks.", "model": "claude-haiku-5-5"}
++{"work_id": 3090906110, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.35, "rationale": "Combinatorial analysis of reaction systems' transition graphs; theoretical computer science with unclear fit, closest to computation theory.", "model": "claude-haiku-5-5"}
++{"work_id": 3174631184, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "Fixed point constructions in tilings and cellular automata simulating Turing machines; computability and dynamics of tilings.", "model": "claude-haiku-5-5"}
++{"work_id": 7134076622, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Operations research survey of offshore wind farm optimization models and solution methods; closest CS category is algorithms/combinatorial optimisation.", "model": "claude-haiku-5-5"}
++{"work_id": 4387356133, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.8, "rationale": "PSPACE-completeness proof for the universal game Turning Tiles.", "model": "claude-haiku-5-5"}
++{"work_id": 2567311606, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "Randomness notions via infinite time Turing machines and effective descriptive set theory; computability theory with logic focus.", "model": "claude-haiku-5-5"}
++{"work_id": 7125235431, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Geoscience study using PLS regression on drill-core hyperspectral data; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2787143214, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.55, "rationale": "Decidability of the weak monadic theory of successor, relating to automata and computability theory; logic-oriented.", "model": "claude-haiku-5-5"}
++{"work_id": 7162015013, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Management and education survey study (SEM-PLS on digital marketing and leadership); not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 4361029879, "categories": ["ml_theory"], "primary": null, "confidence": 0.5, "rationale": "Distributed control theory for large-scale linear systems; closest to control/optimization guarantees but not a clear CS category; weak fit for any listed category.", "model": "claude-haiku-5-5"}
++{"work_id": 4387624329, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Informer-based transformer variant with sparse attention and distilling for dc-dc converter modeling; architecture design applied to a power electronics task.", "model": "claude-haiku-5-5"}
++{"work_id": 2794110548, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Granular computing and cellular automata formalism for decision modeling; soft computing theory with weak fit to listed categories.", "model": "claude-haiku-5-5"}
++{"work_id": 2990666756, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Internal model control for MIMO control systems; control engineering rather than a listed CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 2888657220, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Adaptive filtering algorithm for bilinear form identification; signal processing, not a listed CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 7221079107, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.9, "rationale": "Explicit affine witness isolation with implications for NP subset P/poly; complexity-theoretic barriers.", "model": "claude-haiku-5-5"}
++{"work_id": 2968152157, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "AI-based interpretation of airborne geophysical data for gold exploration; applied geoscience using ML, closest fit is scientific computing.", "model": "claude-haiku-5-5"}
++{"work_id": 4414170577, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Safety-critical control of underactuated vehicle (hovercraft) with control barrier functions; control of physical systems fits robotics.", "model": "claude-haiku-5-5"}
++{"work_id": 2905468286, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Laboratory equipment handbook content for environmental microbiology; not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 3110879399, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.93, "rationale": "Algebraic complexity: VNP, Permanent, circuit size of equations, hardness results.", "model": "claude-haiku-5-5"}
++{"work_id": 4396536343, "categories": ["computer_architecture"], "primary": null, "confidence": 0.0, "rationale": "Radar signal processing for MIMO parameter estimation; not clearly a CS category, and the closest fit (computer_networks/speech_audio) is not warranted. Returned empty for conservatism.", "model": "claude-haiku-5-5"}
++{"work_id": 4288109977, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.35, "rationale": "Corrosion modelling with cellular automata is a computational simulation of a physical process, but the abstract is minimal.", "model": "claude-haiku-5-5"}
++{"work_id": 4417095717, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.35, "rationale": "Proposes a new computational paradigm to reinterpret Transformers; theoretical framing of architectures, with questionable rigor.", "model": "claude-haiku-5-5"}
++{"work_id": 2955693445, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "Bounds on absolute root separation of integer polynomials; closer to number-theoretic/algebraic complexity bounds than algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 4409570626, "categories": ["quantum_computing"], "primary": null, "confidence": 0.3, "rationale": "Probabilistic cellular automata as discrete quantum systems; physics-oriented rather than computing-oriented, so low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 2888883998, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.7, "rationale": "Descriptional complexity and hierarchies of automata and languages; automata theory falls under computational_complexity.", "model": "claude-haiku-5-5"}
++{"work_id": 7214073407, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Finite equational bases for semirings relates to equational logic and finite basis problems, closest to formal reasoning about algebraic theories.", "model": "claude-haiku-5-5"}
++{"work_id": 2809671584, "categories": [], "primary": null, "confidence": 0.0, "rationale": "Number-theoretic experimental study of prime comparisons; pure mathematics, not computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 4206034737, "categories": ["computer_vision"], "primary": null, "confidence": 0.3, "rationale": "Machine learning with remote sensing for air temperature estimation; environmental application, weakly fits computer_vision via satellite data.", "model": "claude-haiku-5-5"}
++{"work_id": 6967003509, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.95, "rationale": "Characterizes lossy catalytic space complexity classes and separations; pure complexity theory.", "model": "claude-haiku-5-5"}
++{"work_id": 2738145410, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Optimization formulations for two-stage stochastic MIPs; convexification and cuts for integer programs, closest to algorithms/combinatorial optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 2952324823, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Numerical simulation and linear stability analysis of a packed-bed chemical reactor model.", "model": "claude-haiku-5-5"}
++{"work_id": 2962975699, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.8, "rationale": "Complexity classification of VCSPs over infinite domains, tractability and NP-hardness.", "model": "claude-haiku-5-5"}
++{"work_id": 7169788258, "categories": ["computer_vision", "speech_audio"], "primary": "computer_vision", "confidence": 0.55, "rationale": "YOLO-based detection on time-frequency spectrogram images for radar signal recognition; signal domain may also fit speech_audio-adjacent work but primarily image detection.", "model": "claude-haiku-5-5"}
++{"work_id": 2995807397, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.8, "rationale": "Proof of strong NP-hardness for a clustering problem with size constraints.", "model": "claude-haiku-5-5"}
++{"work_id": 4409702353, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Survey study of MOOC adoption in health informatics education using UTAUT; social science rather than computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7148239571, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Book on cryptography and Kolmogorov complexity; abstract is minimal, but topic links to complexity theory and cryptography.", "model": "claude-haiku-5-5"}
++{"work_id": 3086506771, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Algebraic study of ordered semihypergroups, a mathematics topic rather than computer science.", "model": "claude-haiku-5-5"}
++{"work_id": 7154114167, "categories": ["hci"], "primary": null, "confidence": 0.3, "rationale": "Fuzzy Mamdani classifier for IoT-based aquaculture water quality; applied fuzzy logic with IoT sensors, weak fit to any category, so conservatively low-confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 4382237584, "categories": ["robotics", "ai_agents"], "primary": "robotics", "confidence": 0.6, "rationale": "Autonomous vehicle decision-making for pedestrian interaction at street crossings; AV as robotic system, with intention-aware agent decision-making.", "model": "claude-haiku-5-5"}
++{"work_id": 4313639087, "categories": ["ai_agents", "ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Networked multi-agent control with PID tracking and stability analysis; control theory rather than learning or AI agents, so low confidence and loosely fits ml_theory or agents.", "model": "claude-haiku-5-5"}
++{"work_id": 2978699180, "categories": ["computational_biology", "nlp"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Automated model construction from scientific literature via NLP for complex biological/ecological networks with simulation and model checking; mixed framing.", "model": "claude-haiku-5-5"}
++{"work_id": 4416034050, "categories": ["ai_agents", "ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Security of LLM-based multi-agent systems: attacks and defenses on MAS roles and topologies.", "model": "claude-haiku-5-5"}
++{"work_id": 4387171855, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Heuristic improvement for numeric AI planning via bounds extraction for LM-cut; heuristic search algorithms.", "model": "claude-haiku-5-5"}
++{"work_id": 7202147256, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Preference-based reward modeling for multi-agent RL to shape emergent behavior.", "model": "claude-haiku-5-5"}
++{"work_id": 7171217691, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Social science and tourism labour commentary on agentic AI; not a computer science research contribution.", "model": "claude-haiku-5-5"}
++{"work_id": 4281774473, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Feature extraction and selection for efficient time series classification (data mining); the abstract is mostly acknowledgements, so classification is tentative.", "model": "claude-haiku-5-5"}
++{"work_id": 7212269269, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.45, "rationale": "Update admission auditing for continual embodied agents with policy updates and replay; RL-centred continual learning with embodied robot context.", "model": "claude-haiku-5-5"}
++{"work_id": 4225668571, "categories": ["nlp", "hci"], "primary": "hci", "confidence": 0.5, "rationale": "Chatbots as conversational agents for learning and coaching, focusing on user interaction and educational systems.", "model": "claude-haiku-5-5"}
++{"work_id": 3095072217, "categories": ["ai_agents", "cloud_computing"], "primary": "ai_agents", "confidence": 0.3, "rationale": "Multi-agent simulation platform for urban mobility during COVID-19; not clearly CS core category. Multi-agent system is the main technique, but application is transport planning; low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 2968942120, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Framework for publishing interconnected open linked data from heterogeneous sources, semantic web data integration.", "model": "claude-haiku-5-5"}
++{"work_id": 7130252400, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Business and marketing study of AI adoption by small businesses; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2911791709, "categories": [], "primary": null, "confidence": 0.6, "rationale": "E-learning course development for storage administration training; educational rather than CS research contribution, no clear category fit.", "model": "claude-haiku-5-5"}
++{"work_id": 4412987259, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Data-driven model-free control of multi-agent consensus with ESO and Lyapunov stability; control theory, closest to ML-free control theory, low fit to ontology.", "model": "claude-haiku-5-5"}
++{"work_id": 4289360054, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Minimax and alpha-beta game AI for Gonu game; game-playing search, closest category is agents/planning for acting.", "model": "claude-haiku-5-5"}
++{"work_id": 3128896073, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Information systems literature review methodology for theory building; not in the ontology's CS categories.", "model": "claude-haiku-5-5"}
++{"work_id": 3037034904, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Transportation policy simulation of NYC transit reopening; transportation engineering rather than CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 3014047513, "categories": ["ml_theory"], "primary": null, "confidence": 0.2, "rationale": "Controllability of multiagent systems via lambda-matrix; systems and control theory, not a clear ontology fit.", "model": "claude-haiku-5-5"}
++{"work_id": 3122776204, "categories": [], "primary": null, "confidence": 0.2, "rationale": "Containment control of nonlinear multi-agent systems; control theory, no clear ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 7077881861, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Multi-agent RL in mixed-motive games with gradient adjustment and reward restructuring; core contribution is an RL optimization method.", "model": "claude-haiku-5-5"}
++{"work_id": 7160854959, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Control theory for multi-agent systems with adaptive NN and event-triggering; control engineering rather than CS categories, weakly fits none clearly.", "model": "claude-haiku-5-5"}
++{"work_id": 3126546068, "categories": ["ai_agents", "computer_architecture"], "primary": "ai_agents", "confidence": 0.35, "rationale": "Multi-agent system framework for power distribution network protection and self-healing; power systems domain with agent-based control.", "model": "claude-haiku-5-5"}
++{"work_id": 2997077628, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Automated negotiation ranking of alternatives by multi-agent systems, with construction domain test case.", "model": "claude-haiku-5-5"}
++{"work_id": 3035359449, "categories": ["ai_agents", "computational_biology"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Multiagent methods applied to EEG brain activity recognition; agent-based analysis of neural data.", "model": "claude-haiku-5-5"}
++{"work_id": 4212806673, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Reaction-diffusion modeling of Dictyostelium signaling pattern formation; biological modeling with numerical simulation.", "model": "claude-haiku-5-5"}
++{"work_id": 4416750122, "categories": ["ai_agents", "reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Decentralized multi-agent RL for mixed traffic control with robot vehicles at intersections.", "model": "claude-haiku-5-5"}
++{"work_id": 2945626893, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Explainable-AI-based generative policy models for autonomous coalition systems at the edge.", "model": "claude-haiku-5-5"}
++{"work_id": 2969980661, "categories": ["ai_agents", "computer_networks"], "primary": "computer_networks", "confidence": 0.5, "rationale": "Diffusion protocol for link failure detection and resource utilization in networked multi-agent systems.", "model": "claude-haiku-5-5"}
++{"work_id": 4402920523, "categories": ["ml_theory"], "primary": null, "confidence": 0.3, "rationale": "Distributed consensus control for discrete-time multi-agent systems; control theory, weakly fits ai_agents or none.", "model": "claude-haiku-5-5"}
++{"work_id": 7137881325, "categories": ["nlp", "llm"], "primary": "llm", "confidence": 0.3, "rationale": "Essay on AI (prompt engineering) in high-school ENEM writing education; not clearly CS research, weakly fits LLM use in education. Conservative low confidence.", "model": "claude-haiku-5-5"}
++{"work_id": 7218192222, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Educational use of NetLogo multi-agent simulations to teach complex systems properties; primarily education research, not clearly CS category.", "model": "claude-haiku-5-5"}
++{"work_id": 2898995053, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Literature study of game characteristics of electronic negotiations in negotiation support systems; IS/HCI-oriented.", "model": "claude-haiku-5-5"}
++{"work_id": 3037046004, "categories": ["computer_vision", "hci"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Collaborative satellite image analysis tool with retrainable neural networks and human-AI feedback loops for humanitarian mapping.", "model": "claude-haiku-5-5"}
++{"work_id": 4416453808, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.8, "rationale": "LLM-based multi-agent system design framework with task decomposition and controller-mediated subtasks.", "model": "claude-haiku-5-5"}
++{"work_id": 4362606359, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Formal framework for causal games combining causal inference and game theory; closest to causal inference/probabilistic models, with game-theoretic aspects.", "model": "claude-haiku-5-5"}
++{"work_id": 3004892764, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Interview study on conversational agents' affordances and effects on customers and employees; user/organisational study.", "model": "claude-haiku-5-5"}
++{"work_id": 4390619203, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Socioeconomic overview of distance education growth in Brazil; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 2898910837, "categories": ["ai_safety", "reinforcement_learning"], "primary": "deep_learning_architectures", "confidence": 0.3, "rationale": "Adaptive fuzzy control of nonlinear multiagent systems with delays; control theory, weakly fits ML categories. Low confidence, primary set to a best fit.", "model": "claude-haiku-5-5"}
++{"work_id": 7219983900, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Information science study of Brazilian land tenure information regime; not computer science research.", "model": "claude-haiku-5-5"}
++{"work_id": 7132739348, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.8, "rationale": "Defeasible reasoning and KLM-style logics for conditional entailment is formal knowledge representation/logic framed as principled reasoning theory; closest fit is ml_theory, though it is logic-oriented.", "model": "claude-haiku-5-5"}
++{"work_id": 7162579366, "categories": ["autonomous_scientific_discovery", "ai_agents", "llm"], "primary": "autonomous_scientific_discovery", "confidence": 0.9, "rationale": "Autonomous research agent system producing papers with verifiability audits; core is AI-driven scientific process.", "model": "claude-haiku-5-5"}
++{"work_id": 2920799053, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Book chapter on history and analysis of Hex and Go game-playing programs; not clearly within any ontology category.", "model": "claude-haiku-5-5"}
++{"work_id": 7164056265, "categories": ["ai_agents", "llm", "software_engineering"], "primary": "ai_agents", "confidence": 0.8, "rationale": "Design patterns for compound multi-agent LLM systems with tool agents in an engineering workflow.", "model": "claude-haiku-5-5"}
++{"work_id": 7153320825, "categories": ["ai_agents", "generative_ai"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Multi-agent AI system automating documentary video production pipeline; involves video generation/editing.", "model": "claude-haiku-5-5"}
++{"work_id": 4415035946, "categories": ["ai_agents", "llm", "scientific_computing"], "primary": "ai_agents", "confidence": 0.7, "rationale": "LLM-based cooperative agents for explainable property prediction in quantum chemistry and materials science.", "model": "claude-haiku-5-5"}
++{"work_id": 2993634112, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Educational study on mobile learning and computational thinking with students; closest is HCI/learning technology, weakly fits.", "model": "claude-haiku-5-5"}
++{"work_id": 7154382442, "categories": ["ai_agents", "reinforcement_learning", "llm"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Multi-agent RL training of LLM agents for social deduction with reward optimization.", "model": "claude-haiku-5-5"}
++{"work_id": 7214219730, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Control-theoretic analysis of finite-time bounded consensus in multi-agent systems; closest fit is theory, though it is systems control rather than ML.", "model": "claude-haiku-5-5"}
++{"work_id": 7124257480, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.9, "rationale": "Cooperative multi-agent RL experience replay and reward decomposition method.", "model": "claude-haiku-5-5"}

modified taxonomy/validation/ratings_claude-sonnet-5-5.jsonl +2190 −0

++{"work_id": 4400648397, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Symbolic computation algorithms for factoring difference operators and solving difference equations.", "model": "claude-sonnet-5-5"}
++{"work_id": 2966152009, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Power electronics converter modelling; electrical engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7216832254, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Education management information system model; not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204908987, "categories": [], "primary": null, "confidence": 0.95, "rationale": "English grammar learning study; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391926769, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Computational thinking in vocational education; education research, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226248671, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Supply chain performance measurement with a balanced scorecard; business/management.", "model": "claude-sonnet-5-5"}
++{"work_id": 2896570120, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.45, "rationale": "Combinatorial construction of complementary sequence sets and codes via a paraunitary generator; signal-processing flavored.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408731690, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Conference proceedings entry on AI in education; no CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281753644, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Multi-criteria decision-making for materials selection; engineering decision method, not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4380303543, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Power system protection and wind turbine control; electrical engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4211201992, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Educational R&D of interactive multimedia for English reading; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4283813889, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Game-theoretic equilibria in signaling games; algorithmic game theory with calculation methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396572044, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Lightweight transformer with contrastive learning for fault diagnosis from multi-sensor data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412353443, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Remote sensing mapping of urban villages using building footprints and nightlight imagery with RF classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219753707, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Community service teacher training in technology literacy; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2806158045, "categories": ["robotics"], "primary": "robotics", "confidence": 0.35, "rationale": "Fuzzy logic control of a gantry crane system, a control/mechatronic problem; closest category is robotics.", "model": "claude-sonnet-5-5"}
++{"work_id": 2982518672, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.5, "rationale": "Proposes a wireline communication scheme encoding information in noise statistics; communication system design.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204116048, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Systematic review of digital educational platforms in basic education; education research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112863481, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Multimedia teaching in physical education; education-focused, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3176660608, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Planetary/geoscience spectral unmixing of minerals; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206870115, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education/communication essay on digital teaching; not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409833556, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Cultural literacy infographic design; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990710531, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control engineering of ball suspension in an air pipeline; mechanical/control, not clearly robotics or CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386952809, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Survey study of gaming motives and well-being; user-focused social computing, weak CS fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225947117, "categories": [], "primary": null, "confidence": 0.85, "rationale": "School library management; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3039944072, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Engineering program accreditation relevance; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389567417, "categories": [], "primary": null, "confidence": 0.8, "rationale": "English pronunciation teaching using an app; education research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3109176755, "categories": [], "primary": null, "confidence": 0.55, "rationale": "Pure combinatorics/mathematics of Kakeya sets; no algorithmic framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128287921, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "AI tools in textile design education, student engagement; human-AI interaction, weak CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3093363583, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Educational mobile app for children; interactive system development, mostly education.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035934964, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Mathematical optimization theory on complementarity problems and game equilibrium; closest fit is algorithms (combinatorial optimisation and game theory), though only loosely.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399338544, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Radar target tracking with weighted least squares; signal processing and not clearly in any CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388262188, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Millimeter-wave radar displacement measurement is instrumentation and signal processing, not a CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 2967228095, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Survey sampling statistics estimators; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285385149, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Education study on teacher digital competence; not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4302287363, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Social science study of a literacy activity; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393541793, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Archaeological viewshed analysis; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388202433, "categories": [], "primary": null, "confidence": 0.9, "rationale": "EFL teaching qualitative study; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3200409248, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Clinical neuroimaging study; no computational method contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3007072699, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.4, "rationale": "Portable device classifying heart sound with cepstral features and a supervised classifier; audio signal processing, but biomedical.", "model": "claude-sonnet-5-5"}
++{"work_id": 3007925741, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Ultrasonic flowmeter measurement physics and FE simulation; instrumentation, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387771625, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education/psychology study of teacher candidates' stereotypes; not CS research per se.", "model": "claude-sonnet-5-5"}
++{"work_id": 4297802412, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education policy analysis, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7139130213, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Industrial engineering cycle-time modelling of AS/RS with Petri nets; operations research rather than CS focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385550355, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Cultural/humanities analysis of a video game; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409248967, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Sensor array sampling for magnetocardiography; instrumentation/biomedical engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394836987, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Pure mathematics on fractional differential equations existence; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3193522400, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Preschool art curriculum education; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3098057909, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Sociological study of African virtual communities; social science, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413096664, "categories": [], "primary": null, "confidence": 0.75, "rationale": "CSCL applied to college English teaching; education research, not CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166265487, "categories": [], "primary": null, "confidence": 0.9, "rationale": "A course catalog entry for a capstone course; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3197510543, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education research on a dropout intervention in an online course; learning analytics but main contribution is educational, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4360976908, "categories": [], "primary": null, "confidence": 0.92, "rationale": "English-language education survey on TikTok use; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215297435, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational development of online lessons; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281735920, "categories": [], "primary": null, "confidence": 0.92, "rationale": "Education article on presentation competencies; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3006201666, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Vocational education R&D of a mobile learning app; the contribution is educational.", "model": "claude-sonnet-5-5"}
++{"work_id": 4291313249, "categories": [], "primary": null, "confidence": 0.93, "rationale": "Literature review on distance education in Brazil; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219459307, "categories": [], "primary": null, "confidence": 0.93, "rationale": "Survey of e-learning use in business education; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2803291160, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Web science analysis of social media user responses to health videos; social computing study.", "model": "claude-sonnet-5-5"}
++{"work_id": 4236001208, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Statistics textbook chapter on time series ARMA models; not CS research per se.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402341510, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Expert system using Dempster-Shafer reasoning for medical diagnosis; closest is automated reasoning/expert system, weak fit.</br>", "model": "claude-sonnet-5-5"}
++{"work_id": 2957642544, "categories": [], "primary": null, "confidence": 0.8, "rationale": "EFL language-learning education study, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281729487, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Children's batik coloring book design; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281615972, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Survey of nursing education IT readiness; education research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415542380, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Literature review on reading literacy and communication technology; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3116015520, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Law education distance learning; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396949355, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Device-level FinFET noise characterization at cryogenic temperature; relates to VLSI device reliability.", "model": "claude-sonnet-5-5"}
++{"work_id": 3201351346, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "CMOS rectifier circuit design for RF energy harvesting; low-power VLSI design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2891299477, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Dielectric elastomer energy generator physics/mechatronics, not computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3087361831, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Chemical engineering process control of absorption column; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3081174150, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Choosing big data processing methods for e-learning; a weak fit to big-data processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3207299996, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Geophysical satellite anomaly analysis for earthquakes; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390807490, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Editorial history of a journal; not CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3138970505, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Educational data analysis of online learning behavior; education research, not core CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3212108115, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Education survey on student projects; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389301359, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Educational measurement with IRT on student data; education research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3120906435, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Demographic assessment of geolocated Twitter data for migration; social science focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4249995757, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Particle swarm metaheuristic for multi-objective microgrid scheduling.", "model": "claude-sonnet-5-5"}
++{"work_id": 3159049567, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Library services survey study; library management, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3175870648, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Library science thesis on Web 2.0 communication; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133374282, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education study on digital competence of students; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2994564899, "categories": [], "primary": null, "confidence": 0.9, "rationale": "General article on ICT in education; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405519900, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Media studies of Chilean telenovela industry; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403910223, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Literature review on teacher professional development with AI; education research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285808510, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Mobile game for language learning; an interactive system, but the abstract gives no details.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367555086, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational quasi-experiment on programming learning; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385356940, "categories": [], "primary": null, "confidence": 0.95, "rationale": "English language teaching case study; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4377195986, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control theory for 2D Markov jump systems; engineering, not within the CS ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413172752, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Educational study of video media and learning outcomes; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312554546, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.55, "rationale": "Circuit design for piezoelectric energy harvesting with SC DC-DC converters; low-power VLSI/circuit design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2994665920, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Soil moisture sensing with fiber optics is environmental/hydrology measurement, not a CS topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 3184466402, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education studies on digital educational resources in Kazakhstan and Russia; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203518940, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Session abstract on AI in higher education teaching; pedagogy and policy rather than CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4298631416, "categories": [], "primary": null, "confidence": 0.85, "rationale": "E-learning effectiveness case study in higher education management; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402526116, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Scoping review of data and AI literacy; library and information science and education, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169646709, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Book on organizational complexity management; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4280616584, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Education 4.0 and regional development; education policy, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394974551, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Legal paper on criminal law and cybercrime; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7201179105, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Probabilistic stochastic-matrix model for uncertain reasoning chains, a probabilistic modelling contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3152238796, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Education/psychology study on smartphone and social-media interference with reading comprehension; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148607934, "categories": ["robotics"], "primary": "robotics", "confidence": 0.4, "rationale": "Autonomous train alerting of vehicles at level crossings; a safety system for autonomous rail, loosely robotic/vehicle systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387754658, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Pure mathematics on variational inequality stability in Hilbert spaces; not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2943056810, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Nonlinear circuit dynamics of coupled oscillators; not clearly a CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402215387, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Philosophy/humanities guide to Edgar Morin; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3036170140, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Design and development of an e-commerce information system using the Waterfall method.", "model": "claude-sonnet-5-5"}
++{"work_id": 3005366057, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Practical tips for data automation in data journalism; no clear CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3212504661, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Process sensor technology roadmap for the process industry; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133133565, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Educational philosophy essay on AI in teaching; not technical CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035309673, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education literature review of educational technology; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3205297713, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Editorial on standards ecosystem; not CS research in a specific category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3159931479, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Survey study of employees' stress and coping around after-hours instant-messaging use; social computing and user behavior, though it is largely an information-systems or management study.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166124027, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Urban photovoltaic potential estimation using GIS; energy research, not core computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220147160, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Television episodic structure and curriculum; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133386248, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Retrospective on institutional repositories network, with no abstract; library administration, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2889050697, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Mixed logical dynamical modelling and optimal scheduling of multi-energy systems; control and power engineering, not clearly in any CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390251033, "categories": [], "primary": null, "confidence": 0.9, "rationale": "No substantive information; not identifiable as CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215366142, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Librarians' communication role; library science, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4376526549, "categories": [], "primary": null, "confidence": 0.97, "rationale": "English teaching materials development; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388721237, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Coordinated pricing for coupled power and transportation networks via equilibrium models; electrical and transportation engineering, not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413905440, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Teacher training for university teachers in Algeria, including distance learning; education research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407389586, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Optical metrology of small angles via speckle imaging; physics/optics rather than a CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4311440779, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Business management survey study on digital transformation and environmental performance; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4322706772, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Electrical machine design; electrical engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399577834, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Language education study on WebQuest; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4294309008, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Career/profession piece with no abstract; no clear CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387450145, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Community service training on Canva for school; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2924637482, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Qualitative communication/media study on Instagram infographics; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388692834, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Consortium description; no research contribution in a CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378804495, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Affective computing: personality recognition from ECG with VR emotional stimuli.", "model": "claude-sonnet-5-5"}
++{"work_id": 2899278391, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "ML-based ECG fragmented QRS detection; biomedical/health informatics method.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215205522, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Community service training on Excel for report cards; educational ICT training, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2988437044, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Educational R&D of learning videos; not CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3217539441, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "User-centric perception study on ICT-enabled learning environment design; borderline HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 7172412385, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Pure number theory/quaternion algebra result, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410790364, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Vocational education training base construction; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367663421, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Engineering management quality evaluation of remanufactured products; neural net only a parameter estimator.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413050241, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.65, "rationale": "ML prediction of preterm birth from genetic and lifestyle data; biomedical informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 3137938609, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Knowledge management and digital libraries in education research; library science, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215521419, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Community service introducing Scratch to schoolchildren; education activity, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2899602491, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Information systems effect on elderly adherence; health-oriented human-technology study, weak HCI fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4383695424, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Analytic number theory (L-function universality), not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 2890545363, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Mathematical ecology/epidemiology bifurcation model; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2921387549, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education survey on talent competencies; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411385469, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Airport security organisation study, not computational research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4220744985, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Language education study using animation; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385809862, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Health policy/data-sharing epidemiological study, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405194088, "categories": ["nlp"], "primary": "nlp", "confidence": 0.5, "rationale": "Digital corpus processing and lexical analysis with NLP of historical texts.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406402165, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Solar plant production vs simulation software comparison; engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4235935708, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Mathematical model of nonlinear oscillations in a MEMS generator; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410359890, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Development economics report; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4395448276, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Review of AI in education using TAM; education research, not core CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127568039, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Student survey on digital teaching; education/public health, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3034138045, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational management scenario study in Egypt; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381193254, "categories": [], "primary": null, "confidence": 0.98, "rationale": "Philosophy of psychology; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220079635, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Qualitative education study of AI in language apps; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3213996293, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Optical sensor physics/instrumentation; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990253946, "categories": ["edge_computing", "algorithms"], "primary": "edge_computing", "confidence": 0.4, "rationale": "IoT-based evacuation system with evolutionary algorithm for deployment of information posts; IoT system plus metaheuristic.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203847693, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Edited volume on language education technology pedagogy; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413000914, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Survey on teacher stress and AI literacy; psychology/education.", "model": "claude-sonnet-5-5"}
++{"work_id": 2997518615, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Meta-synthesis of digital competence studies; education research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221210578, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Training workshop on an online learning platform in Indonesian education; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405738901, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Nonlinear physics of chimera states in neuron models; computational neuroscience framing but mainly physics, not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4210484898, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Manufacturing/drilling engineering; compressed sensing is incidental.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382986001, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Review of open access publishing fees; scholarly communication, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2946426731, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Crowdfunding and film industry business study; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386047869, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.4, "rationale": "ML regression on hyperspectral imaging for fat content in food; image analysis application, borderline.", "model": "claude-sonnet-5-5"}
++{"work_id": 7216070232, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Editorial on ethics in nutrition science; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387426085, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Education study on pesticide detection talent cultivation; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3168268628, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "CSCW-venue web tool for collaborative team formation using forgetting curves.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405513290, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Physics of synchronization transitions in adaptive simplicial complexes; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390074168, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Management case study of manufacturing digital transformation; not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2947916735, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Testing/assessment model for e-learning diagnostics; weakly interactive systems, mostly education.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213988930, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library science/education philosophy; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112250358, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Solar resource estimation; energy/meteorology, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148627082, "categories": ["algorithms", "computer_vision"], "primary": "algorithms", "confidence": 0.6, "rationale": "Bio-inspired mixed-variable metaheuristic (GWO) applied to image classification and hyperspectral unmixing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7218427777, "categories": [], "primary": null, "confidence": 0.85, "rationale": "School library services study; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4352989561, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Information diffusion model on social networks with SIR-based modelling; closest fit is algorithmic modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128218809, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Language learning with messaging apps; education research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406799739, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.35, "rationale": "Variational inequality existence theory and iterative algorithm; mathematics/optimization.", "model": "claude-sonnet-5-5"}
++{"work_id": 3129513804, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Reflective essay on computing education research; not technical CS and no matching category.", "model": "claude-sonnet-5-5"}
++{"work_id": 2915662082, "categories": [], "primary": null, "confidence": 0.6, "rationale": "CS education program poster; no ontology category fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 3006452720, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Spatiotemporal interpolation/imputation model for geospatial data; numerical method with ELM ensemble.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388634849, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Qualitative study of women's social media use and agency in ICT4D; social computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4205387857, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Statistical time series analysis of climate data; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162195214, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Random forest on small student dataset; applied educational data analysis, no clear CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7112126402, "categories": [], "primary": null, "confidence": 0.45, "rationale": "General overview of AI use in education; no specific category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3212702011, "categories": [], "primary": null, "confidence": 0.55, "rationale": "Photodiode device characterization; optoelectronics, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226049650, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Optics textbook chapter; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409157433, "categories": [], "primary": null, "confidence": 0.55, "rationale": "Econophysics stock correlation networks; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3080490474, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Optical nondestructive testing; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411680001, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Explores integrating AI tools into children's creative art activities; human-AI interaction in education, weakly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312292089, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Qualitative education study of an after-school programme; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412583047, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Educational program evaluation in primary schools; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391250710, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Engineering education on online labs; pedagogy rather than CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2954250517, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Library system workflow session; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3129713058, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational technology experiment on learning strategy; not CS research proper.", "model": "claude-sonnet-5-5"}
++{"work_id": 3206918036, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Quality assessment of remote sensing land products; geoscience validation, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412965306, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Proposes LSTM-BiLSTM model for infrastructure condition estimation from vibration signals; application of recurrent networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415671687, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Geostatistical simulation to improve ML classification with uncertainty quantification; probabilistic methodology applied to geology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404412038, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Empirical career analysis of CIOs; information systems/management, not core CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3202598135, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Spectrum allocation in wireless networks using a bio-inspired metaheuristic optimization algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 2969566816, "categories": ["computer_architecture", "computer_networks", "cybersecurity"], "primary": "computer_architecture", "confidence": 0.55, "rationale": "FPGA OpenFlow switch architecture with built-in DDoS defence functions for SDN.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411949771, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.92, "rationale": "RPL routing extension for QoS in IoT low-power lossy networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409370970, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Transformer-based intrusion detection for DDoS in 5G networks; the main contribution is security detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117461702, "categories": ["privacy_preserving_computation", "computer_networks"], "primary": "privacy_preserving_computation", "confidence": 0.65, "rationale": "Privacy-protecting sparse data compression and collection in WSNs with perturbation and lightweight encryption.", "model": "claude-sonnet-5-5"}
++{"work_id": 2890181437, "categories": ["computer_networks", "cybersecurity"], "primary": "computer_networks", "confidence": 0.55, "rationale": "Satellite telecommand link with CCSDS synchronization and anti-jamming for availability.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399801188, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.75, "rationale": "Outage performance analysis of fluid-antenna backscatter wireless communication.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162436281, "categories": ["cybersecurity", "applied_cryptography"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Offline smart lock authentication using BLE and Ed25519-signed presentations; a security system design.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204457438, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Zero Trust framework with ML-based malicious node detection for 6G IoT networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 2982187970, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "UAV-aided backhaul traffic scheduling in RAN with path planning.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011142689, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.85, "rationale": "5G synchronisation as a service, SDN/NFV; networking platform.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413557542, "categories": ["computer_networks", "robotics"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Analytical latency model of DDS reliable delivery over UDP in ROS 2; networking performance modelling for robot middleware.", "model": "claude-sonnet-5-5"}
++{"work_id": 2887023911, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Broadcast transmission scheme (LDM, OFDM, LDPC) evaluated by simulation; physical-layer, closest to networking.", "model": "claude-sonnet-5-5"}
++{"work_id": 7217270607, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "AI for threat detection in network security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414744692, "categories": ["federated_learning", "computer_networks", "edge_computing"], "primary": "federated_learning", "confidence": 0.75, "rationale": "Learned digital over-the-air aggregation for federated edge learning with a learned decoder.", "model": "claude-sonnet-5-5"}
++{"work_id": 3202853167, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "ARQ retransmission and index coding analysis for reliable broadcast.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400315068, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "ML-based anomaly detection for IoT network security.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138862310, "categories": ["reinforcement_learning", "computer_networks"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Bandwidth estimation for real-time video communication using a new offline RL algorithm; deployed in Teams.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382248155, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.88, "rationale": "In-band telemetry and data plane programmability for network monitoring and operation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3137488241, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Downgrade attack analysis of WPA3 transition mode.", "model": "claude-sonnet-5-5"}
++{"work_id": 4291165406, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.35, "rationale": "Mobile sensor network for air pollution monitoring; mostly environmental sensing, weak fit to networks, but sensor network framing is the closest category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400276720, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.9, "rationale": "5G FR2 mobility and beamforming system-level simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417252163, "categories": ["generative_ai", "computer_networks"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Energy-based diffusion model with MH sampling for MIMO channel estimation; contribution is on the generative model, applied to wireless.", "model": "claude-sonnet-5-5"}
++{"work_id": 3157147353, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.93, "rationale": "802.11 WLAN TXOP scheduling scheme for congestion.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402187505, "categories": ["computer_networks", "nlp"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Intent-driven routing in power backbone networks using LSTM+CRF intent translation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7112597741, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "DNSSEC study and DNS cache poisoning MITM attack.", "model": "claude-sonnet-5-5"}
++{"work_id": 3039350537, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Buffer-aware scheduling in UAV relay sensor networks, with NP-completeness and heuristic.", "model": "claude-sonnet-5-5"}
++{"work_id": 7202339439, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.65, "rationale": "ML detection of spectrum sensing data falsification attacks in cognitive radio networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385988957, "categories": ["computer_networks", "reinforcement_learning"], "primary": "computer_networks", "confidence": 0.85, "rationale": "RL-based routing for TSN time-aware shaper scheduling.", "model": "claude-sonnet-5-5"}
++{"work_id": 3160907559, "categories": ["computer_networks", "software_engineering"], "primary": "software_engineering", "confidence": 0.45, "rationale": "Systematic review of IoT architecture challenges; the abstract is truncated.", "model": "claude-sonnet-5-5"}
++{"work_id": 4290991003, "categories": ["generative_ai", "computer_networks", "privacy_preserving_computation"], "primary": "generative_ai", "confidence": 0.7, "rationale": "GAN-based synthetic network trace generation with fidelity and privacy concerns, for networking tasks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4210382415, "categories": ["federated_learning", "distributed_computing"], "primary": "federated_learning", "confidence": 0.6, "rationale": "Coded communication reduction for distributed multi-task learning with users training locally and a server, which is federated-style learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4375957559, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.5, "rationale": "STAR-RIS covert wireless communication with beamforming optimization, framed as a wireless networking and physical-layer problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404627904, "categories": ["speech_audio", "multimodal_ai"], "primary": "speech_audio", "confidence": 0.4, "rationale": "Target signal extraction via blind source separation guided by multimodal clues, inspired by auditory attention. Applied to wireless interference cancellation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389101774, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Wearable wireless breath monitoring system, i.e. wearable computing/sensing, though mostly sensor hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416894542, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Conceptual framework for network architecture with SDN/NFV for Internet performance.", "model": "claude-sonnet-5-5"}
++{"work_id": 2775587204, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Educational survey of students' attitudes toward social networking sites; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3013810395, "categories": ["edge_computing", "reinforcement_learning", "computer_networks"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Vehicular edge computing with computation offloading and data scheduling solved with deep RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386918894, "categories": ["reinforcement_learning", "computer_networks"], "primary": "computer_networks", "confidence": 0.65, "rationale": "RAN slicing resource allocation in vehicular networks using multi-agent RL (MADDPG).", "model": "claude-sonnet-5-5"}
++{"work_id": 2806094603, "categories": ["ai_accelerators", "computer_networks"], "primary": "ai_accelerators", "confidence": 0.55, "rationale": "GPU acceleration of SVM training and classification, applied to network traffic classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 3207037104, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Survey of attacks and countermeasures in wireless sensor networks; security is the main focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408441962, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Attention-based neural network for malicious traffic detection in IoT networks, which is intrusion detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 2922268566, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Editorial on networking technologies such as SDN, SDA and SD-WAN.", "model": "claude-sonnet-5-5"}
++{"work_id": 2921781262, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.88, "rationale": "Reduces duplicate requests to the SDN controller and the OpenFlow communication load.", "model": "claude-sonnet-5-5"}
++{"work_id": 7107949444, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.85, "rationale": "Measurement and performance analysis of a campus Wi-Fi network.", "model": "claude-sonnet-5-5"}
++{"work_id": 3176443371, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.6, "rationale": "Cache updating strategy minimizing age of information, solved as a constrained Markov decision problem; networking and caching framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3115846872, "categories": ["computer_networks", "robotics"], "primary": "computer_networks", "confidence": 0.6, "rationale": "V2X data dissemination of 3D maps for GNSS positioning, with a data rate analysis of the 802.11p link.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381805382, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.82, "rationale": "Survey of Sybil attack countermeasures in underwater sensor networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 2800766972, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.88, "rationale": "Latency of OpenFlow rule changes in SDN data centre networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410810312, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.7, "rationale": "Outage and error-rate analysis of RIS-assisted multihop wireless networks, supported by a deep learning estimator.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220552095, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Text mining and sentiment analysis of reviews with lexicon and LDA; classical NLP.", "model": "claude-sonnet-5-5"}
++{"work_id": 4283810249, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Span extraction and question generation for low-resource art domain; NLP task.", "model": "claude-sonnet-5-5"}
++{"work_id": 2909157681, "categories": ["nlp", "information_retrieval"], "primary": "nlp", "confidence": 0.6, "rationale": "Tag embeddings and semantic similarity for folksonomies using word embeddings and LDA; text mining.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124304832, "categories": ["nlp"], "primary": "nlp", "confidence": 0.95, "rationale": "POS tagging comparison on a new Pashto corpus; low-resource NLP.", "model": "claude-sonnet-5-5"}
++{"work_id": 4221152634, "categories": ["nlp"], "primary": "nlp", "confidence": 0.95, "rationale": "German dataset for summarization and simplification; NLP resource.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160969555, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Czech valency lexicon, a computational linguistics resource.", "model": "claude-sonnet-5-5"}
++{"work_id": 3033351879, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.7, "rationale": "Task-oriented dialogue with a modified transformer; dialogue task with architecture tweaks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4353004366, "categories": ["llm", "deep_learning_architectures"], "primary": "llm", "confidence": 0.65, "rationale": "Long-context processing by reusing pretrained LMs (SLED), evaluated on SCROLLS; efficient inference for language models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399355922, "categories": ["nlp", "information_retrieval"], "primary": "nlp", "confidence": 0.6, "rationale": "Document matching for chatbots comparing distance measures and BERT; text matching.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399657664, "categories": ["nlp", "database_systems"], "primary": "nlp", "confidence": 0.55, "rationale": "Benchmark for entity summarization over knowledge graphs; NLP resource and evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167593003, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.6, "rationale": "Depression detection from social media text using an LLM-backbone mixture-of-experts; the task framing is text classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387323905, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.7, "rationale": "Unified seq2seq LLM pretraining for table tasks, studied across model scales.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213562850, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Lexical data representation model for classifying text descriptions, using ontologies and hypervector algebra.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166888984, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.5, "rationale": "System demo for reasoning attribution in diagram question answering; the abstract is minimal, so this is inferred from the title.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203972941, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "BERT-based Arabic fake news sentiment analysis, a text classification task.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166865552, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.65, "rationale": "SemEval system comparing prompting, RAG and fine-tuning of GPT-4.1-mini for abductive event reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7212985666, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.7, "rationale": "Feasibility study of LLMs for courtroom machine translation and LLM-as-a-judge evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414404922, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Large multilingual dataset and pipeline for speech generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4377372007, "categories": ["llm", "ai_agents"], "primary": "llm", "confidence": 0.75, "rationale": "Framework in which LLMs self-correct through tool-interactive critiquing, which reduces hallucination.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390215039, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "CCG-based compositional semantics for questions, with a QA evaluation dataset; computational linguistics.", "model": "claude-sonnet-5-5"}
++{"work_id": 2964776615, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Neural question generation from passages is a classic NLP task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412888626, "categories": ["llm", "ai_agents"], "primary": "llm", "confidence": 0.7, "rationale": "Studies ToM representations in LLaMA models and uses them to align conversational responses.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389523964, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.8, "rationale": "Proposes a feature-based knowledge distillation method for compressing language models.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035035809, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.5, "rationale": "Box embeddings for hierarchical and multi-relational knowledge such as WordNet, a representation learning problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285254984, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Ontology-based educational ecosystem for railway universities; not core computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396653207, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.8, "rationale": "Thesis on sentence embeddings and the role of sentence structure in neural language models.", "model": "claude-sonnet-5-5"}
++{"work_id": 347871603, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Framework and guidelines for evaluating machine translation systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417356542, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Benchmarks CNNs against prompt-tuned BERT for sentiment analysis, a text classification task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4361006333, "categories": ["nlp"], "primary": "nlp", "confidence": 0.5, "rationale": "Genetic-algorithm feature selection with classical classifiers for text-based failure prediction, using NLP preprocessing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4212987749, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Sentiment analysis of movie reviews using classical supervised ML models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387389341, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "Text classification of Japanese interview transcripts with morpheme selection and classical ML models.", "model": "claude-sonnet-5-5"}
++{"work_id": 2997143856, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Sentiment classification with RNNs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396695606, "categories": ["deep_learning_architectures", "nlp"], "primary": "deep_learning_architectures", "confidence": 0.55, "rationale": "Noise-based regularization of LSTM/BiLSTM models, evaluated on NLP tasks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406975371, "categories": ["llm", "deep_learning_architectures"], "primary": "llm", "confidence": 0.75, "rationale": "Proposes a context-aware mechanism with attention modulation layers for large language models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404782050, "categories": ["llm", "ai_safety"], "primary": "llm", "confidence": 0.8, "rationale": "Inference-time alignment of LLMs using value-guided decoding.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213318585, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Book on sentiment analysis in NLP, covering text-based and multimodal methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200154996, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "Sentiment and topic classification of COVID-19 news articles (text mining).", "model": "claude-sonnet-5-5"}
++{"work_id": 4225411551, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "SemEval structured sentiment analysis with a BART-based generative approach.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403127614, "categories": ["nlp"], "primary": "nlp", "confidence": 0.65, "rationale": "E-commerce customer service chatbot built with NLP and ML (dialogue).", "model": "claude-sonnet-5-5"}
++{"work_id": 2971319154, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Grammatical error correction with synthetic erroneous data generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226489197, "categories": ["nlp"], "primary": "nlp", "confidence": 0.93, "rationale": "Low-resource named entity recognition with a dual learning framework is a classical NLP information extraction task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285298945, "categories": ["nlp"], "primary": "nlp", "confidence": 0.92, "rationale": "Abusive comment detection in Tamil is low-resource text classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399923297, "categories": ["nlp", "hci"], "primary": "nlp", "confidence": 0.5, "rationale": "Chatbot implementation using NLP, with personality-based interaction and user experience evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4213262540, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Sentiment analysis and topic modeling of English and Tamil tweets.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405737917, "categories": ["nlp"], "primary": "nlp", "confidence": 0.92, "rationale": "Naive Bayes sentiment classification of Indonesian tweets.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169949244, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Abstract gives only the degree programme, so the paper's contribution cannot be judged and it cannot be placed in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7217753392, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.6, "rationale": "Review of embedding fusion methods for NLP and ML, including neural network-based fusion.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400512809, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Sentiment and emotion analysis of Indonesian X data using supervised ML.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204475324, "categories": ["hci"], "primary": "hci", "confidence": 0.75, "rationale": "Qualitative interview study of how students use AI tools, a user study of human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408546512, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Automatic document summarisation using a recurrent neural network.", "model": "claude-sonnet-5-5"}
++{"work_id": 2964302064, "categories": ["hci", "computational_biology"], "primary": "hci", "confidence": 0.4, "rationale": "EEG/MEG fusion for motor imagery BCI classification; BCI sits closest to HCI, with a biomedical signal-processing angle.", "model": "claude-sonnet-5-5"}
++{"work_id": 2946097570, "categories": ["hci"], "primary": "hci", "confidence": 0.85, "rationale": "User-centered requirements elicitation for learning analytics dashboards with stakeholders.", "model": "claude-sonnet-5-5"}
++{"work_id": 3127843396, "categories": ["hci"], "primary": "hci", "confidence": 0.85, "rationale": "Interaction model for mixed-reality interactive narrative experiences.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140640315, "categories": ["computer_vision", "multimodal_ai", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Action recognition using multimodal fusion and adaptive graph convolution, with a VQ-VAE for missing data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131851078, "categories": ["hci", "ai_safety"], "primary": "hci", "confidence": 0.5, "rationale": "AI-enabled PE learning framework with holographic display, gamification and XAI feedback; mainly an interactive system.", "model": "claude-sonnet-5-5"}
++{"work_id": 3196036446, "categories": ["hci"], "primary": "hci", "confidence": 0.9, "rationale": "Assets-based design of socio-technical systems with a vulnerable community; social computing and design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2948072807, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Microwave circuit design (phase shifter) is electrical engineering and falls outside the ontology categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 3039601072, "categories": ["hci"], "primary": "hci", "confidence": 0.92, "rationale": "Field study on audio-mediated reminiscence for blind people; accessibility and interaction design.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171785185, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Clinical systematic review of eye-tracking in infants, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312560958, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "AR-based educational media development and evaluation with students; AR user-focused application.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403160367, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.5, "rationale": "Exploratory agent used to evaluate procedurally generated game levels; agent-based AI-driven game design evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312735085, "categories": ["hci"], "primary": "hci", "confidence": 0.9, "rationale": "New method for evaluating long-term user experience with children, a user study in HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135288805, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Sociolinguistic study of Facebook's effect on Algerian youth writing; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412960450, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Survey of vocational students' perceptions of augmented reality technology; AR user perception study.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220333334, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Design criteria for e-commerce sites; business/web design commentary without clear CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409620177, "categories": [], "primary": null, "confidence": 0.7, "rationale": "English-language education survey of a teaching platform; educational research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203798403, "categories": ["hci"], "primary": "hci", "confidence": 0.92, "rationale": "Conceptual framework on UX challenges in VR training systems: usability, comfort, presence.", "model": "claude-sonnet-5-5"}
++{"work_id": 2980067317, "categories": [], "primary": null, "confidence": 0.4, "rationale": "BCI motor imagery classification using signal features and SVM; biomedical signal processing, with no ontology category as its main contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3166564139, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Routine web-based information system implementation for a university department; no clear research contribution in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413023786, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Systematic review of design principles and outcomes for mobile learning apps; interaction design focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4317438351, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "VR educational game; abstract is empty but the title indicates a VR application for learning, so HCI/VR.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403626708, "categories": ["hci"], "primary": "hci", "confidence": 0.55, "rationale": "Review of AR's effects on consumer behavior with a conceptual model; borderline for CS, closest to XR/HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 2951084085, "categories": ["distributed_computing", "computational_biology"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "PDkit data toolkit for Parkinson's sensor data, scaled with Apache Beam dataflow stream processing; combines scalable dataflow processing with health informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131915600, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.8, "rationale": "UI modernization with frameworks and microservices using an agile methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417220792, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Critical analysis of interdisciplinary research culture between HCI/interaction design and behavioural science for behaviour change design.", "model": "claude-sonnet-5-5"}
++{"work_id": 7198871956, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Qualitative sociolinguistic study of language use on WhatsApp; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7202300623, "categories": ["hci"], "primary": "hci", "confidence": 0.85, "rationale": "Immersive VR data visualization design space with prototypes and a user study.", "model": "claude-sonnet-5-5"}
++{"work_id": 2790900283, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Framework for virtual cognitive experiments using immersive VR and EEG to study user cognition in virtual geographic environments.", "model": "claude-sonnet-5-5"}
++{"work_id": 4221152344, "categories": ["hci", "ai_safety"], "primary": "hci", "confidence": 0.8, "rationale": "Large-scale user study of human-AI teamwork in hiring, examining model bias transfer to human decisions.", "model": "claude-sonnet-5-5"}
++{"work_id": 7119905577, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Questionnaire study of presence, immersion and satisfaction in MR cultural tourism; user experience of XR.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405928278, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Systematic literature review of metaverse research in business; the metaverse is an XR/HCI topic and the venue is an HCI journal.", "model": "claude-sonnet-5-5"}
++{"work_id": 3186159243, "categories": ["hci"], "primary": "hci", "confidence": 0.55, "rationale": "Pilot study of VR for teaching in schools, covering user experience and learning with immersive technology.", "model": "claude-sonnet-5-5"}
++{"work_id": 7112876774, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Study of XR headsets for code-tracing lessons using eye gaze and motivation questionnaires; a user study of immersive interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225091482, "categories": ["hci"], "primary": "hci", "confidence": 0.92, "rationale": "CHI paper on an open-ended sensor kit for self-tracking, a personal informatics and ubiquitous computing study.", "model": "claude-sonnet-5-5"}
++{"work_id": 2911964379, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Game-theoretic incentive algorithms for crowdsourcing quality control based on zero-determinant strategies; algorithmic game theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285312412, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Media and cultural studies essay on TikTok videos; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413901110, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Educational survey of student perceptions of AI in language classes; pedagogy rather than computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7123361757, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Experiment on emotional responses to 360\u00b0 VR video, covering VR and affective responses.", "model": "claude-sonnet-5-5"}
++{"work_id": 3216674623, "categories": ["hci"], "primary": "hci", "confidence": 0.88, "rationale": "Wearable assistive communication device prototype for deaf-blind users; accessibility and wearable computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161768941, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Sociolinguistic analysis of address terms on Twitter/X; linguistics rather than computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413188996, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education essay on digital divide and teacher training; not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3189825209, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Human-centered and critical design case study on communicating user insights for smart home products; design research within HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400638471, "categories": ["hci"], "primary": "hci", "confidence": 0.55, "rationale": "User-centered design guide for low-tech artefacts presented at a cognitive ergonomics conference; HCI design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3138774926, "categories": ["hci"], "primary": "hci", "confidence": 0.85, "rationale": "Empirical user study of eye, mouse and scroll behavior in web browsing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7216627434, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Community service training on AR learning media for math teachers; education-focused, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400230898, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Wireless intracortical BCI chip/system with on-chip spike detection and compression, an embedded circuit design paper at ISCAS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124922965, "categories": ["hci"], "primary": "hci", "confidence": 0.92, "rationale": "Position paper on annotation interaction techniques in tactile graphics for blind users; accessibility.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200159390, "categories": ["hci"], "primary": "hci", "confidence": 0.45, "rationale": "Overview of systematic reviews on digital mental-health interventions; borderline, closest to HCI and health technology.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125427419, "categories": ["hci"], "primary": "hci", "confidence": 0.88, "rationale": "Thematic review and framework for incentive design in human-AI decision-making studies.", "model": "claude-sonnet-5-5"}
++{"work_id": 7132825323, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Survey experiment on digital human likeness, uncanny valley and purchase intention; human-AI interaction and affective responses.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117881138, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Adaptive weight aggregation in FL for multi-institution polyp segmentation; FL method plus medical image segmentation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4317566642, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.65, "rationale": "Density estimation for generating disclosure-protected synthetic micro-data; statistical method with a privacy goal.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388516975, "categories": ["privacy_preserving_computation", "computer_vision"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "Homomorphic encryption of CT images combined with deep learning classification for private lung cancer diagnosis.", "model": "claude-sonnet-5-5"}
++{"work_id": 7110219066, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.9, "rationale": "New noise distribution for smooth-sensitivity differential privacy mechanisms.", "model": "claude-sonnet-5-5"}
++{"work_id": 7123282485, "categories": ["privacy_preserving_computation", "ai_safety", "ai_agents"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "LLM agents re-identify interview participants, a deanonymization attack on released data, with misuse and mitigation implications.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313598458, "categories": ["privacy_preserving_computation", "ml_theory"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Causal inference from locally differentially private data, with minimax optimality results.", "model": "claude-sonnet-5-5"}
++{"work_id": 4360995097, "categories": ["federated_learning", "privacy_preserving_computation"], "primary": "federated_learning", "confidence": 0.88, "rationale": "Gradient inversion attack on federated learning that recovers private inputs.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171902819, "categories": ["privacy_preserving_computation", "hci"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Differential privacy for EEG community sensing, with noise induced at acquisition by RF signals.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405184264, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.85, "rationale": "Theoretical privacy-utility analysis of re-identification risk in the Topics API.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404037745, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.85, "rationale": "Symmetric private information retrieval over a noisy channel with colluding servers, an information-theoretic analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3139240920, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.9, "rationale": "K-anonymity data anonymization proof of concept.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226227596, "categories": ["cybersecurity", "privacy_preserving_computation"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "ML-based privacy scoring of Android apps from permissions; mobile security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4318812474, "categories": ["privacy_preserving_computation", "deep_learning_architectures", "federated_learning"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Patch shuffling on transformers to protect data in split learning, with formal privacy guarantees and defenses against attacks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411949626, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.88, "rationale": "Prototype-based FL under domain heterogeneity on edge devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214618239, "categories": ["cybersecurity", "llm"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Adversarial fraud transaction generation with LLMs to benchmark fraud detection and catastrophic forgetting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405710613, "categories": ["privacy_preserving_computation", "deep_learning_architectures"], "primary": "privacy_preserving_computation", "confidence": 0.4, "rationale": "Book on graph learning with emphasis on privacy preservation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165521315, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Optimal mechanism design for data brokers selling market segmentation; algorithmic game theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 3137146239, "categories": ["privacy_preserving_computation", "computer_networks"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Secret-sharing-based identity broadcast for BLE contact tracing to reduce privacy exposure.", "model": "claude-sonnet-5-5"}
++{"work_id": 2971047472, "categories": ["privacy_preserving_computation", "computational_complexity", "formal_verification"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Differential privacy distances on labelled Markov chains, with complexity results and bisimulation distances.", "model": "claude-sonnet-5-5"}
++{"work_id": 7168322783, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Federated few-shot learning for IoMT AFib detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393952232, "categories": ["privacy_preserving_computation", "robotics"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "System using secure two-party computation to protect citizen location privacy around delivery drones.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133196926, "categories": ["privacy_preserving_computation", "deep_learning_architectures"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Differentially private training algorithms preserving gradient sparsity for large embedding models.", "model": "claude-sonnet-5-5"}
++{"work_id": 2899391355, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Empirical study of GDPR data portability requests; privacy regulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2945715100, "categories": ["privacy_preserving_computation", "edge_computing"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Privacy system for smart service/IoT platforms limiting disclosed private information.", "model": "claude-sonnet-5-5"}
++{"work_id": 2886835905, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "Hash-based attribute anonymization scheme for location privacy in LBS continuous queries.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125647931, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Economics/public administration discussion of AI use, not clear CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3093013316, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.4, "rationale": "Legal analysis of privacy and public health law conflicts; weak CS content, privacy regulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3012891874, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Graph-based statistical analysis of identity attributes and identity theft risk.", "model": "claude-sonnet-5-5"}
++{"work_id": 3004671875, "categories": ["privacy_preserving_computation", "cybersecurity"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Location privacy attack on geosocial networks inferring user locations.", "model": "claude-sonnet-5-5"}
++{"work_id": 2998582260, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Mimic learning shares student model instead of sensitive health data to preserve privacy.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415149235, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.35, "rationale": "ML-based classification of GPS and check-in location data; spatial data mining with a privacy motivation, but not a clear fit to any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3217252216, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.85, "rationale": "Synthesis of Gaussian privacy mechanisms that minimize mutual information against inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401845216, "categories": ["federated_learning", "cybersecurity"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Review of privacy and security of federated learning in UAV systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403764267, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Healthcare data privacy protection and countermeasures against data leakage.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155107746, "categories": ["privacy_preserving_computation", "generative_ai"], "primary": "privacy_preserving_computation", "confidence": 0.5, "rationale": "Selecting synthetic data for transmission in distributed settings, using optimal control; synthetic data as a privacy tool.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382045981, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Studies business resilience under GDPR regulation, which falls under privacy regulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388951965, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Social network analysis of the PGP web of trust, concerning key trust and authentication.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386204720, "categories": ["privacy_preserving_computation", "software_engineering"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Log redaction of sensitive information using source code analysis and data flow, for software diagnosis.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165783687, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Survey study of user trust, privacy concerns and purchase intention toward GenAI-personalized advertising; human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 7123336800, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.9, "rationale": "New exponential mechanism formulation for differential privacy that incorporates priors.", "model": "claude-sonnet-5-5"}
++{"work_id": 2904199664, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "Anonymizing co-owned photos in social networks to limit privacy loss is a privacy-engineering contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416037029, "categories": ["federated_learning", "llm"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Personalized federated LoRA fine-tuning of LLMs, with aggregation weights optimized by hierarchical Bayesian optimization.", "model": "claude-sonnet-5-5"}
++{"work_id": 2952484830, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "(k,p)-anonymity algorithm for sanitizing transactional databases.", "model": "claude-sonnet-5-5"}
++{"work_id": 3204605723, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.35, "rationale": "Agent-based simulation of pandemic response using wearables with privacy-preserving contact chains; privacy is only part of the framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405979621, "categories": ["cybersecurity", "privacy_preserving_computation"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Cancelable biometric template generation protects biometric authentication and privacy; this falls under biometric security and template protection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385248498, "categories": ["hci", "privacy_preserving_computation"], "primary": "hci", "confidence": 0.7, "rationale": "Consumer behavior study of privacy concerns and resistance to voice assistants, framed through HCI theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412164648, "categories": ["federated_learning", "privacy_preserving_computation"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Client selection in federated learning under personalized differential privacy budgets.", "model": "claude-sonnet-5-5"}
++{"work_id": 2786217712, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.5, "rationale": "Covers the legislative history of encryption and privacy regulation (GDPR, UNHRC), which falls under privacy regulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118160301, "categories": ["federated_learning", "privacy_preserving_computation", "ai_safety"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Privacy-preserving federated evaluation and statistical testing of cross-client performance fairness for FL models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386192152, "categories": [], "primary": null, "confidence": 0.5, "rationale": "A rule-based expert system for chicken disease diagnosis using forward chaining is an applied system with no clear fit to an ontology category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405359480, "categories": ["blockchain", "cybersecurity"], "primary": "blockchain", "confidence": 0.8, "rationale": "Blockchain smart-contract access control for IoT data sharing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011050248, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.4, "rationale": "ML on cheap sensors to detect IoT device co-location; IoT sensing system with weak fit to categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 2900068116, "categories": ["computer_networks", "algorithms"], "primary": "computer_networks", "confidence": 0.55, "rationale": "Participant selection in mobile crowdsensing with energy-efficient transmission, formulated as a CMDP.", "model": "claude-sonnet-5-5"}
++{"work_id": 3090747742, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Manufacturing reliability modelling; mostly industrial engineering, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4214663994, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.3, "rationale": "Wearable sensor data fusion for medical IoT using ML; weak fit and a retraction notice.", "model": "claude-sonnet-5-5"}
++{"work_id": 3163035554, "categories": ["edge_computing", "algorithms"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Task offloading and scheduling in wireless-powered MEC using a swarm metaheuristic.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117133120, "categories": ["reinforcement_learning", "federated_learning"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Federated transfer PPO for home energy management.", "model": "claude-sonnet-5-5"}
++{"work_id": 3211844447, "categories": ["edge_computing", "computer_networks"], "primary": "edge_computing", "confidence": 0.5, "rationale": "Software-defined smart factory fire protection spanning cloud/fog/edge with runtime reconfiguration.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410068037, "categories": ["cybersecurity", "hci"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Survey study of IoT security awareness among students; human factors of security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413558883, "categories": ["applied_cryptography", "edge_computing"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Lightweight CP-ABE with cooperative computation offloading for constrained IoT devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 4383427890, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.6, "rationale": "mmWave positioning with AoA estimation error analysis for IoT; wireless networking/localization framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7126134868, "categories": ["reinforcement_learning", "cybersecurity", "ai_safety"], "primary": "reinforcement_learning", "confidence": 0.55, "rationale": "Multi-agent actor-critic with Markov games for malware defense, plus SHAP explainability.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389319530, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Conference preface with no research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3007216950, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Persian general popular overview of IoT with no clear research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 3025242379, "categories": ["algorithms", "computer_networks"], "primary": "algorithms", "confidence": 0.55, "rationale": "Online reverse-auction incentive mechanisms with provable properties for mobile crowdsensing; algorithmic game theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385153790, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Agricultural pest counting with a sensor testbed and standard ML; main contribution is not clearly CS research in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281393175, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.9, "rationale": "Computation offloading and resource allocation in MEC, optimized with a metaheuristic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312555758, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Wearable IoT remote monitoring system for patient prehabilitation; ubiquitous and wearable computing application.", "model": "claude-sonnet-5-5"}
++{"work_id": 3096106062, "categories": ["edge_computing", "deep_learning_architectures"], "primary": "edge_computing", "confidence": 0.8, "rationale": "On-device TinyML for embedded and IoT units, using pruning and approximation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3199685485, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.7, "rationale": "Low-power sensor readout interface circuit with a switching regulator for IoT; analog and VLSI design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3216809321, "categories": ["applied_cryptography", "cloud_computing"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Proposes an advanced RC6 encryption algorithm to secure sensitive IoT data in a private cloud.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401883192, "categories": [], "primary": null, "confidence": 0.4, "rationale": "General overview chapter on ML in IoT with no clear core contribution in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390284844, "categories": ["edge_computing", "blockchain", "federated_learning"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Resource allocation for blockchain-based federated learning on MEC servers, with offloading and optimisation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400095687, "categories": ["edge_computing", "blockchain", "cloud_computing"], "primary": "edge_computing", "confidence": 0.55, "rationale": "Architecture combining cloud, fog, edge and blockchain for SDG data governance.", "model": "claude-sonnet-5-5"}
++{"work_id": 3176753072, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Survey of IoT security requirements, architecture and challenges.", "model": "claude-sonnet-5-5"}
++{"work_id": 3184417063, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Contactless load monitoring via modality distillation with an oracle-apprentice network; this is sensing plus knowledge distillation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390905002, "categories": ["edge_computing", "algorithms"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Joint model deployment and task offloading in vehicular edge computing, with an NP-hardness proof and an approximation algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 7156021111, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Soil moisture prediction with a swarm-optimised SVM on IoT agricultural data is an application with weak CS category fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399525972, "categories": ["edge_computing", "deep_learning_architectures", "reinforcement_learning"], "primary": "edge_computing", "confidence": 0.8, "rationale": "End-edge collaborative inference with model partitioning and offloading, using DRL and a fuzzy neural network structure.", "model": "claude-sonnet-5-5"}
++{"work_id": 2886809270, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Lightweight stream-cipher-based scheme with dynamic keys and mutual authentication for IoT.", "model": "claude-sonnet-5-5"}
++{"work_id": 4289520433, "categories": ["edge_computing", "reinforcement_learning"], "primary": "edge_computing", "confidence": 0.88, "rationale": "Dependency-aware task offloading in vehicular edge computing using a DDPG reinforcement learning algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367394805, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.4, "rationale": "Architectural overview of an IoT-based smart healthcare system; IoT platform as a computing substrate, weak fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164347165, "categories": ["federated_learning", "sustainable_computing", "edge_computing"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Energy-aware federated learning framework for edge devices aiming at green computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378364725, "categories": ["blockchain", "cybersecurity"], "primary": "blockchain", "confidence": 0.6, "rationale": "Literature review on integrating blockchain with IoT for security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411194662, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Control-theoretic MPC for cyber-physical systems under DoS attacks; fits no category clearly.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411000650, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.45, "rationale": "Chapter on IoT sensor data and deep learning for early warning systems, with scalability and edge deployment considerations.", "model": "claude-sonnet-5-5"}
++{"work_id": 3047548318, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Survey of fog computing in the IoT environment: architecture, advantages and applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281719112, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Electronic nose gas sensor hardware prototype; sensing hardware not in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410347576, "categories": ["edge_computing", "reinforcement_learning"], "primary": "edge_computing", "confidence": 0.88, "rationale": "Deep Q-learning-based dynamic task offloading and resource management in fog computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3097229291, "categories": ["reinforcement_learning", "edge_computing"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Novel Fisher-information reward function for deep RL sensing policies on energy-constrained IoT devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410459414, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Survey of predictive maintenance in Industry 4.0; an industrial engineering topic with no clear CS category as its main contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 2902217519, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.92, "rationale": "Security evaluation of smart home platforms: access control flaws and privilege escalation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416583070, "categories": ["edge_computing", "privacy_preserving_computation", "federated_learning"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Edge healthcare monitoring framework with privacy-preserving encryption and federated learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 2918201812, "categories": ["edge_computing", "ml_theory"], "primary": "edge_computing", "confidence": 0.55, "rationale": "Location prediction with a factor graph model applied to service migration in mobile edge computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4294691681, "categories": ["edge_computing", "deep_learning_architectures"], "primary": "edge_computing", "confidence": 0.65, "rationale": "Lightweight STFT CNN for ECG classification, quantized and pruned and deployed on an edge device.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415397548, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.45, "rationale": "Systematic review of generative AI and IoT in precision agriculture.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409047392, "categories": ["cybersecurity", "applied_cryptography"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "IoT authentication and data protection using RSA, AES and ECC with variable key sizes.", "model": "claude-sonnet-5-5"}
++{"work_id": 4379398101, "categories": ["edge_computing", "algorithms"], "primary": "edge_computing", "confidence": 0.75, "rationale": "Fog-cloud workload allocation optimizing power and delay via PSO and branch-and-bound.", "model": "claude-sonnet-5-5"}
++{"work_id": 7123925437, "categories": ["edge_computing", "algorithms"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Metaheuristic priority-driven task allocation in vehicular edge computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413824611, "categories": ["edge_computing", "cybersecurity"], "primary": "edge_computing", "confidence": 0.6, "rationale": "Fog-based secure healthcare data management with edge encryption and access control over 5G.", "model": "claude-sonnet-5-5"}
++{"work_id": 4377038089, "categories": ["computer_vision", "cybersecurity"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Survey of face morph attack detection in images; biometric image analysis and security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4362518865, "categories": ["cybersecurity", "reinforcement_learning"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "RL-based black-box adversarial attack against SQL injection detectors/WAFs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401110449, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "GAN-based phishing detection; security application.", "model": "claude-sonnet-5-5"}
++{"work_id": 3090172217, "categories": ["cybersecurity", "applied_cryptography", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Cloud data confidentiality survey proposing hybrid ABE/IBE re-encryption scheme.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171536369, "categories": ["ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.9, "rationale": "Defends LLM safety alignment against neuron-pruning attacks.", "model": "claude-sonnet-5-5"}
++{"work_id": 2952116314, "categories": ["cybersecurity", "blockchain"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Empirical study of DDoS impact on a cryptocurrency exchange.", "model": "claude-sonnet-5-5"}
++{"work_id": 3197514244, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.88, "rationale": "Honeypot design for microservices to collect attack data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219206452, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.4, "rationale": "Socio-legal analysis of cyber crimes in India; only loosely computing-related.", "model": "claude-sonnet-5-5"}
++{"work_id": 3167654854, "categories": ["nlp", "cybersecurity"], "primary": "nlp", "confidence": 0.5, "rationale": "ML model detecting fraudulent used-goods posts using text features.", "model": "claude-sonnet-5-5"}
++{"work_id": 3044959568, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Survey study on human factors in information security policy compliance.", "model": "claude-sonnet-5-5"}
++{"work_id": 2944025354, "categories": ["applied_cryptography", "cloud_computing"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Key sharing and management scheme combining Diffie-Hellman, ECC and OTP for homomorphic encryption in cloud.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390504465, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "EEG motor-imagery biometric authentication using a hybrid pretrained CNN plus classical classifier.", "model": "claude-sonnet-5-5"}
++{"work_id": 3031484172, "categories": ["hci", "cybersecurity"], "primary": "hci", "confidence": 0.8, "rationale": "Mixed-methods user study of smart home device sharing and access control needs.", "model": "claude-sonnet-5-5"}
++{"work_id": 7170084806, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Continuous behavior-based user authentication using an autoencoder and PSO.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200300237, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "News report of a data breach at T-Mobile.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154943096, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.9, "rationale": "ML-based phishing URL detection with a Random Forest.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412549228, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.75, "rationale": "Chaotic image encryption combined with hash-LSB steganography.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392971839, "categories": ["llm", "generative_ai", "reinforcement_learning", "ai_safety"], "primary": "llm", "confidence": 0.75, "rationale": "Model-level watermarking of LLM text trained via RL co-training of detector and LLM to track misuse.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167720828, "categories": ["privacy_preserving_computation", "cybersecurity"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Security framework for privacy and data protection in outsourced data, with key generation, masking, access control and compliance.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416409576, "categories": ["generative_ai", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Multimodal deepfake and fake news detection combining text, image and audio.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214148734, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Black-box adversarial evasion attack on MalConv malware detector using memetic algorithm; malware/security of AI systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285320914, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.9, "rationale": "Digital forensics tools for information security.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171969272, "categories": ["generative_ai", "computer_vision", "cybersecurity"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Deepfake image detection with hybrid VGG16-CNN; AI-generated content detection, also vision and security.", "model": "claude-sonnet-5-5"}
++{"work_id": 3083517685, "categories": ["cybersecurity", "hci", "privacy_preserving_computation"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Usable, privacy-preserving, fraud-resistant digital identity for smartphones; authentication, usability, privacy.", "model": "claude-sonnet-5-5"}
++{"work_id": 4298144048, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.92, "rationale": "Comparison of web vulnerability scanners for penetration testing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4310682878, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Android malware and search rank fraud detection on Google Play.", "model": "claude-sonnet-5-5"}
++{"work_id": 7011534987, "categories": ["cybersecurity", "database_systems"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Digital forensics phone recording analysis via sequential pattern mining with ASP; forensics and data mining.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131118580, "categories": ["cybersecurity", "ml_theory"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Web attack detection with uncertainty-aware ensemble deep kernel learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7200539267, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.4, "rationale": "Software for secure information transmission; abstract is only a citation, so classification is based on the title.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385221456, "categories": ["cybersecurity", "nlp", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Voice phishing detection from text with a hybrid CNN-BiLSTM-attention model.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393303862, "categories": ["cybersecurity", "software_engineering"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Threat modeling practices for CPS, interview studies of practitioners, and automated threat identification; security plus software engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4239390297, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Deep learning applied to malicious URL detection; primarily a security task.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990181665, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "Overview of protective security; weak abstract but security-themed.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154473294, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Comparison of conventional and explainable AI threat analysis in SIEM systems; security operations focus, with XAI as a secondary theme.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215892371, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.88, "rationale": "Comparative analysis of Android malware detection methods, including neural network approaches.", "model": "claude-sonnet-5-5"}
++{"work_id": 2816623457, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "DoS attack detection in cloud environments using a dendritic cell mechanism; intrusion detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4380988124, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Cybersecurity architecture and standards for smart buildings with IoT.", "model": "claude-sonnet-5-5"}
++{"work_id": 3162129019, "categories": ["cybersecurity", "quantum_computing"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Exploiting radiation-induced bit-flips as an attack; also considers soft-error consequences in quantum computers.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112992601, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "Survey study of cybersecurity risk management in companies during COVID-19; human and organisational security factors.", "model": "claude-sonnet-5-5"}
++{"work_id": 7001137095, "categories": ["cybersecurity", "computer_networks"], "primary": "cybersecurity", "confidence": 0.65, "rationale": "IKEv2 remote access VPN configuration and performance evaluation, covering security and networking.", "model": "claude-sonnet-5-5"}
++{"work_id": 4289655646, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.93, "rationale": "RLWE-based post-quantum certificateless authenticated key exchange with a security proof in the eCK model.", "model": "claude-sonnet-5-5"}
++{"work_id": 3089767340, "categories": ["hci", "cybersecurity"], "primary": "hci", "confidence": 0.7, "rationale": "Sonic feedback for password strength is an interaction design problem; it is also tied to authentication security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411782417, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Multi-class malware classification framework for cloud environments, using ML for malware detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155614231, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Survey study on the efficacy of social engineering prevention measures, which is a human-factors security topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 2995203890, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.5, "rationale": "DWT-based image watermarking for robustness and security, which is close to multimedia security and information hiding.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409103867, "categories": ["nlp", "deep_learning_architectures"], "primary": "nlp", "confidence": 0.7, "rationale": "Spam text classification with an LSTM and attention ensemble, which is a text classification task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393105534, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Interactive educational media on sex education for teens, which is art and education rather than CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409248207, "categories": ["cybersecurity", "applied_cryptography"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Lightweight continuous authentication for IoT under zero trust, built on identity-based encryption.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403677606, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Risk analysis methodology for GenAI systems, validated on prompt injection attacks against LLMs.", "model": "claude-sonnet-5-5"}
++{"work_id": 3041436169, "categories": ["cybersecurity", "deep_learning_architectures"], "primary": "cybersecurity", "confidence": 0.75, "rationale": "RNN-based classification of power grid contingencies and cyber-attacks for intrusion detection, with embedded deployment.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160546262, "categories": ["software_engineering", "ai_agents", "algorithms"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Software remodularization/architecture recovery framed as multi-agent negotiation with formal protocol proofs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4311440938, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.45, "rationale": "Web app hosting platform design using Nginx and Python; engineering/deployment contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409591325, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Agile methodologies in IT transformation, project management and leadership.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128328612, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Thesis on designing and implementing an online store on PrestaShop CMS.", "model": "claude-sonnet-5-5"}
++{"work_id": 6986525772, "categories": ["formal_verification", "blockchain", "programming_languages", "software_engineering"], "primary": "blockchain", "confidence": 0.7, "rationale": "Pre-deployment analysis of smart contracts, static analysis for data races, and DSL design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200040010, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Tutorial chapter on Python unittest test automation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3157577106, "categories": ["software_engineering", "algorithms"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Software effort estimation with COCOMO parameters optimized by PSO hybrid with ML models.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118011917, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.3, "rationale": "Introductory programming book chapter on functions; minimal content.", "model": "claude-sonnet-5-5"}
++{"work_id": 2972326490, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Design of an Android event-information app with web service using SDLC Waterfall.", "model": "claude-sonnet-5-5"}
++{"work_id": 2932905330, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.9, "rationale": "Tool support for test smell summaries improving test suite quality awareness.", "model": "claude-sonnet-5-5"}
++{"work_id": 3008788199, "categories": ["software_engineering", "nlp"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Classifies security requirements in specifications using ontology and ML; requirements engineering with text classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387781545, "categories": ["software_engineering", "hci"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Structured editor extension with domain-specific representations of code; developer tooling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401907809, "categories": ["software_engineering", "reinforcement_learning"], "primary": "software_engineering", "confidence": 0.85, "rationale": "Testing autonomous driving systems using RL adversarial agents; ICST testing paper.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410538665, "categories": ["llm", "software_engineering", "ai_safety"], "primary": "llm", "confidence": 0.8, "rationale": "Benchmark evaluating correctness consistency and robustness of code-generating LLMs.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219410480, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "IT project management methodology for developing an application.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166522127, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.65, "rationale": "White-label software reuse approach for e-commerce mobile apps.", "model": "claude-sonnet-5-5"}
++{"work_id": 3143264371, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.9, "rationale": "Quality assurance framework for software process and product based on ISO standards.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392404200, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.75, "rationale": "Ensemble ML model predicting software project failure; empirical SE project risk.", "model": "claude-sonnet-5-5"}
++{"work_id": 4253254765, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Excel tutorial book chapter, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4310841548, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.85, "rationale": "Automatic REST API code generator using templates; development automation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386014351, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Construction-industry digital procurement and business process study; not core computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4280590664, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Agile business transformation management framework, loosely tied to software deployment and agile development.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312839447, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Book chapter on building Android apps; developer tooling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206446312, "categories": ["ai_agents", "robotics"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Decentralized multi-agent holonic control architecture for mobile assembly systems in manufacturing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4316021200, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.92, "rationale": "Code smell identification and refactoring with performance evaluation on a mobile app.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388972704, "categories": ["blockchain", "software_engineering"], "primary": "blockchain", "confidence": 0.7, "rationale": "Service composition across heterogeneous blockchains, decentralized applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215194540, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.85, "rationale": "Running Jasmine tests with Maven; software testing tooling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4293177572, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Development of a web management information system for an ISP; application building.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203691417, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.9, "rationale": "Software effort estimation using lazy learning algorithms evaluated on datasets.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135050868, "categories": ["software_engineering", "quantum_computing"], "primary": "software_engineering", "confidence": 0.9, "rationale": "Modifiability scenario for quantum software, based on mining release histories of projects.", "model": "claude-sonnet-5-5"}
++{"work_id": 3044591241, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Development of an Android attendance app; application-building case, weak fit but closest to software engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 7217619139, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Android student monitoring app development using standard design tools; system development.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162132995, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Antibody CDR design with a new architecture for protein design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385395724, "categories": ["software_engineering", "cloud_computing"], "primary": "software_engineering", "confidence": 0.8, "rationale": "SOA and microservices architecture for SME information system with CI/CD and Kubernetes.", "model": "claude-sonnet-5-5"}
++{"work_id": 2908635758, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Mathematics education using Monte Carlo simulation of Parrondo's paradox; not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7113083064, "categories": ["hci"], "primary": "hci", "confidence": 0.55, "rationale": "Qualitative study of student and teacher perspectives on GenAI in programming education; human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 2981150652, "categories": ["hci"], "primary": "hci", "confidence": 0.85, "rationale": "UIST browser extension tool supporting novices in ML prototyping; interactive system.", "model": "claude-sonnet-5-5"}
++{"work_id": 7081987054, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.95, "rationale": "Action research on technical debt management in an IT company.", "model": "claude-sonnet-5-5"}
++{"work_id": 4239998720, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Web-based futsal rental system built with CodeIgniter; application development.", "model": "claude-sonnet-5-5"}
++{"work_id": 3026019741, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Web application for dormitory room data management using the waterfall method.", "model": "claude-sonnet-5-5"}
++{"work_id": 7116117897, "categories": ["llm", "software_engineering"], "primary": "llm", "confidence": 0.8, "rationale": "Benchmark evaluating LLM code reasoning under realistic complexity.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163811937, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Statistical resampling/bootstrap methodology and sensitivity to imbalance.", "model": "claude-sonnet-5-5"}
++{"work_id": 2859484040, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.9, "rationale": "Metamorphic testing to find implementation bugs in ML image classifiers.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154701096, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Web application system design and testing for team collaboration and workflow automation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117153023, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Development of a web monitoring tool using incremental methodology, user stories, microservices and SUS evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2931497058, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Android tour guide app designed with UML and black-box tested; application development.", "model": "claude-sonnet-5-5"}
++{"work_id": 4205450578, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Book chapter on defining generic functions and types, a language feature.", "model": "claude-sonnet-5-5"}
++{"work_id": 4398220511, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.35, "rationale": "Web-based information system development for teacher data management; abstract is only references.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155216780, "categories": ["software_engineering", "privacy_preserving_computation"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Frames testing of machine unlearning as a software QA problem using causal fuzzing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405173371, "categories": ["software_engineering", "information_retrieval"], "primary": "software_engineering", "confidence": 0.85, "rationale": "Information-theoretic evaluation framework for software traceability link recovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4383108538, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.9, "rationale": "Task-aware grasp estimation and pick-and-place on a real robot; core robotics manipulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403919332, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.85, "rationale": "Semantic mapping on a legged robot for search and rescue using vision-based recognition.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410526140, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.75, "rationale": "Educational SCARA robot with vision-based control and CNN/autoencoder fault detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 3182010183, "categories": ["robotics", "computer_networks"], "primary": "computer_networks", "confidence": 0.55, "rationale": "Tethered UAV relay with a high-gain antenna for CNPC links; communication-focused UAV system design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388100051, "categories": ["robotics", "deep_learning_architectures"], "primary": "robotics", "confidence": 0.7, "rationale": "GNN-based vehicle path forecasting for automated driving.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394659963, "categories": ["ai_safety", "robotics"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Chapter on trust, safety, assurance and regulation of robotics and autonomous systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385413558, "categories": ["robotics"], "primary": "robotics", "confidence": 0.9, "rationale": "Robust backstepping interception control for an underactuated USV with experiments.", "model": "claude-sonnet-5-5"}
++{"work_id": 4315632170, "categories": ["ai_safety", "hci"], "primary": "ai_safety", "confidence": 0.4, "rationale": "Ethical analysis of social robots in eldercare from a Confucian perspective, with a governance protocol.", "model": "claude-sonnet-5-5"}
++{"work_id": 3173643949, "categories": ["computer_vision", "robotics"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Visual place recognition via ICM graph optimization over image sequences.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404604190, "categories": ["robotics", "algorithms"], "primary": "robotics", "confidence": 0.7, "rationale": "Kinematic position analysis with a multi-affine interval solver (trapezoid method) for robotics problems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413925424, "categories": ["robotics", "ai_agents"], "primary": "robotics", "confidence": 0.7, "rationale": "Decentralized safe control of multi-robot systems using neural control barrier functions and MPC; robotics with multi-agent flavor.", "model": "claude-sonnet-5-5"}
++{"work_id": 2909162014, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "CNN-based online failure prediction for an industrial robot threading task using force-torque data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415135180, "categories": ["robotics"], "primary": "robotics", "confidence": 0.45, "rationale": "Scenario-based safety certification of connected automated vehicles using UAV data; autonomous vehicle systems focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313471462, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Adaptive MPC path tracking control for autonomous vehicles using a neural network.", "model": "claude-sonnet-5-5"}
++{"work_id": 4319034475, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Fault-tolerant control of an underwater salvage robot with adaptive neural network.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416749351, "categories": ["algorithms", "ai_agents", "robotics"], "primary": "algorithms", "confidence": 0.7, "rationale": "Improves conflict-based search for multi-agent pathfinding with learned conflict prediction; combinatorial search algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214885326, "categories": ["robotics", "ai_agents", "computer_vision"], "primary": "robotics", "confidence": 0.65, "rationale": "Multi-UAV recruitment via motion-encoded messages decoded using RGB and event vision in simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382935921, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control-theoretic design for nonlinear networked systems with delays; control engineering, not clearly in CS ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 2909059777, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Fuzzy cooperative flocking algorithm for multi-robot systems compared against Boids in simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161591331, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "RL framework using online human preference interventions for real-world robot manipulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4297857948, "categories": ["computer_vision", "robotics"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Recovers 3D cable models from a single image with 2D/3D networks, aimed at robot manipulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3034448848, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Fault detection and isolation for a UAS air data system using integrated navigation; closest fit is robotic and aerial systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394880330, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.3, "rationale": "Remote-sensing aerosol optical depth estimation from ground data, validated against satellite AOD; weak fit to remote-sensing analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391947971, "categories": ["robotics"], "primary": "robotics", "confidence": 0.9, "rationale": "Design and testing of a robotic skin with proximity and tactile sensors for human-robot interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4307184398, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education survey of preservice teachers on computational thinking; not CS research in the ontology's sense.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313227226, "categories": ["robotics", "algorithms"], "primary": "robotics", "confidence": 0.6, "rationale": "Multi-UAV coordinated trajectory planning solved with an improved metaheuristic (symbiotic organisms search).", "model": "claude-sonnet-5-5"}
++{"work_id": 4382059457, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Marketing and service-management experiments on perceptions of frontline robots; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410227174, "categories": ["robotics"], "primary": "robotics", "confidence": 0.35, "rationale": "EKF with adaptive filter for non-stationary signals, motivated by robotic control state observers.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403277362, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Learning from demonstration framework for movement primitives, tested on real robot tasks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281717654, "categories": ["robotics"], "primary": "robotics", "confidence": 0.92, "rationale": "Mobile robot localization using AMCL with a semantic grid map and SLAM.", "model": "claude-sonnet-5-5"}
++{"work_id": 2903260265, "categories": ["robotics", "ai_agents"], "primary": "robotics", "confidence": 0.6, "rationale": "Behavior planning under uncertainty for autonomous vehicles using QMDP and MCTS; a robotic planning problem with a decision-theoretic planning flavour.", "model": "claude-sonnet-5-5"}
++{"work_id": 4210424983, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Path planning for an unmanned ship using sensor fusion and a firefly algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 4283759396, "categories": ["robotics"], "primary": "robotics", "confidence": 0.75, "rationale": "Miniature autonomous vehicles as a testbed for autonomous driving methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412030501, "categories": ["robotics"], "primary": "robotics", "confidence": 0.95, "rationale": "Bimanual robot manipulation of deformable food objects.", "model": "claude-sonnet-5-5"}
++{"work_id": 2886313903, "categories": ["robotics"], "primary": "robotics", "confidence": 0.9, "rationale": "Legged locomotion control of an active SLIP model.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011793074, "categories": ["robotics"], "primary": "robotics", "confidence": 0.88, "rationale": "MPC-based path tracking control for autonomous vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367860403, "categories": ["robotics"], "primary": "robotics", "confidence": 0.97, "rationale": "Design of a mobile quad-arm robot with locomotion and manipulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403447662, "categories": ["robotics"], "primary": "robotics", "confidence": 0.95, "rationale": "Tactile in-hand manipulation with a robotic hand.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381429872, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "UAV obstacle avoidance and patrol path planning with sensor fusion.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391695417, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Localization and slip compensation for autonomous robot navigation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3006755905, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Mobile robot with gesture-controlled arm and object detection; robotic system.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160616345, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Offline-to-online RL method with action chunking and theory; applied to robot manipulation benchmarks and VLA models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392494944, "categories": ["robotics", "hci"], "primary": "robotics", "confidence": 0.8, "rationale": "Research summary on human-humanoid collaboration, robot learning and interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 2788120109, "categories": ["robotics", "hci"], "primary": "robotics", "confidence": 0.8, "rationale": "Teleoperation control architecture with adjustable autonomy and usability study.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415293255, "categories": ["robotics"], "primary": "robotics", "confidence": 0.92, "rationale": "Humanoid state estimation via LiDAR-inertial-kinematic odometry (localisation/SLAM).", "model": "claude-sonnet-5-5"}
++{"work_id": 7160828610, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "System identification and pointing control of a robot blimp.", "model": "claude-sonnet-5-5"}
++{"work_id": 3126252149, "categories": ["robotics"], "primary": "robotics", "confidence": 0.8, "rationale": "Sensor-fusion ego-motion estimation and localisation for autonomous race vehicle.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403010033, "categories": ["robotics"], "primary": "robotics", "confidence": 0.85, "rationale": "Multi-robot path planning with DWA and A*, formation and sensor fusion on hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 3130719769, "categories": ["robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "Kinematic chain estimation/calibration of excavator using inertial sensors.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135139717, "categories": ["robotics"], "primary": "robotics", "confidence": 0.9, "rationale": "Real2Sim2Real transformer policy for bimanual robot handover tasks; robot learning and HRI.", "model": "claude-sonnet-5-5"}
++{"work_id": 7158557735, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.55, "rationale": "Probability result on sub-Gaussian variables and convex order; theoretical foundations relevant to learning theory, though borderline for CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410294818, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Regularizer for deep imbalanced regression; a neural training method with no clearer category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215048402, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied comparison of resampling methods on a small breast cancer dataset with WEKA; no clear ontology category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391093173, "categories": ["federated_learning", "privacy_preserving_computation"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Analyzes differential privacy (output perturbation, DP-SGD) in federated learning, with membership inference evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219508654, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Routine application of random forest to classify medical student skills; no category's main contribution applies.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133993797, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Deep clustering combined with Dempster-Shafer evidence theory for uncertain data, with closed-form updates.", "model": "claude-sonnet-5-5"}
++{"work_id": 3179767201, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Deep reject-option classifiers with calibration proofs and generalization bounds.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163158107, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.35, "rationale": "Dropout prediction with interpretability and fairness analysis; mostly applied, fairness and interpretability are a notable element.", "model": "claude-sonnet-5-5"}
++{"work_id": 3155002205, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.8, "rationale": "Krylov subspace methods for partial SVD and rank estimation, which are numerical linear algebra.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160115409, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.88, "rationale": "Online Bayesian conformal prediction with coverage guarantees and regret bounds.", "model": "claude-sonnet-5-5"}
++{"work_id": 3126933325, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "Optimistic Q-learning agent with regret analysis; RL theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4327500503, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Active learning for rare event detection in streams; generic ML with no matching category clearly dominant.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171392614, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied ensemble ML for process control in oil refining; no ontology category fits clearly.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214886155, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Causal discovery with DAG constraints combined with biclustering via optimization; principled ML methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402510141, "categories": [], "primary": null, "confidence": 0.35, "rationale": "Nonmonotonic reasoning inference operators in knowledge representation; no category matches well.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388726461, "categories": [], "primary": null, "confidence": 0.35, "rationale": "Active learning sampling for manufacturing QA; applied ML without matching category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7170116302, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Educational performance prediction with MLP; applied ML with no matching category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385985189, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Ensemble ML for geological mineral prospectivity mapping; geoscience application.", "model": "claude-sonnet-5-5"}
++{"work_id": 7151296083, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Deep anomaly detection in images via pretext tasks, with application to fraud detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4296563198, "categories": ["ml_theory", "ai_safety"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Bayesian feature importance explanations with Dirichlet distribution, tradeoffs with privacy and fairness.", "model": "claude-sonnet-5-5"}
++{"work_id": 7143524691, "categories": ["ai_safety", "generative_ai"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Integrates responsible, explainable and generative AI; focuses on bias, interpretability and governance.", "model": "claude-sonnet-5-5"}
++{"work_id": 4236967208, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Textbook chapter on kernel methods and the kernel trick.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409046524, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Clustering with bandit feedback studies the computation/information gap, which is a learning-theory topic; the abstract is missing, so confidence is moderate.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408365506, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Music education study using basic logistic regression; the contribution is in education, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4305012164, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Bayesian reversible-jump MCMC for mixture models; a probabilistic and statistical method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404916629, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Proposes fuzzy clustering algorithms for relational data; a principled clustering methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 2737706773, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Large-scale benchmark comparing random forest and logistic regression; an ML methodology study.", "model": "claude-sonnet-5-5"}
++{"work_id": 2895972434, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Meta-learning for recommending the combination method of classifier ensembles; general ML methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3126343538, "categories": ["ai_safety", "deep_learning_architectures"], "primary": "ai_safety", "confidence": 0.55, "rationale": "Framework for evaluating the interpretability of deep models by committee consensus.", "model": "claude-sonnet-5-5"}
++{"work_id": 3000034219, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.8, "rationale": "Causal models for dynamical systems; causal inference theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 3012094357, "categories": [], "primary": null, "confidence": 0.4, "rationale": "One-class SVDD anomaly detection for operations data; generic ML with no matching category, so none is assigned.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225697294, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Classifies cloud images using deep neural networks, with imbalanced-data techniques.", "model": "claude-sonnet-5-5"}
++{"work_id": 2897243396, "categories": ["ml_theory", "scientific_computing"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Sparse grid least-squares regression with coordinate rotation; a numerical, kernel-style regression method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409493966, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Detects malicious anomalies in network traffic with ML (Prophet, RNN).", "model": "claude-sonnet-5-5"}
++{"work_id": 7166662462, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.85, "rationale": "Theoretical analysis of early training dynamics of an autoencoder for outlier detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133193233, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.92, "rationale": "Minimax estimation rates for a Markov Gaussian mixture; statistical learning theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392903168, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Diffusion-model-based anomaly detection for multivariate time series.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416873414, "categories": ["llm"], "primary": "llm", "confidence": 0.4, "rationale": "Industrial anomaly detection with LLM-generated explanations; the LLM component is the main novelty.", "model": "claude-sonnet-5-5"}
++{"work_id": 4379538576, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Random forest classifier with penalized discriminant analysis; a principled statistical ML method.", "model": "claude-sonnet-5-5"}
++{"work_id": 2928518543, "categories": ["nlp"], "primary": "nlp", "confidence": 0.55, "rationale": "Unsupervised text-content analysis of Reddit communities for anomaly detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7208751990, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.3, "rationale": "Applied ML financial risk monitoring with stream processing; weak fit to any category, only loosely distributed stream processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385346229, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Bayesian meta-learning with amortized inference and faster convergence; probabilistic ML methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 2900419943, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Accelerated augmented Lagrangian method with iteration-complexity analysis; optimization algorithm convergence.", "model": "claude-sonnet-5-5"}
++{"work_id": 7126019639, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.88, "rationale": "Diffusion-based text-conditioned medical anomaly image synthesis plus downstream detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4234642263, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Routine application of standard ML classifiers to credit card fraud; no clear category contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7132956912, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.88, "rationale": "Rademacher generalization bounds for neural networks with feature diversity and derived regularizers.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213442406, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Multi-objective ensemble hyperparameter selection for imbalanced clinical prediction; biomedical informatics method.", "model": "claude-sonnet-5-5"}
++{"work_id": 3027701685, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Applying ML anomaly detection to APM time series; thin and no clear fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 3173110490, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.85, "rationale": "Online Bayesian nonparametric clustering via variational inference; probabilistic methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396611119, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied ML for PV forecasting in energy domain; no CS category clearly fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 4365509942, "categories": ["formal_verification", "ml_theory"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Classification technique that generates formulas and models for model checking; formal methods used for supervised learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 3204747322, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Process algebra decomposition with bisimulation proofs to reduce state space exploration.", "model": "claude-sonnet-5-5"}
++{"work_id": 3005800578, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.3, "rationale": "Personal account of the move from logic to computational argumentation; weak fit with automated reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 3165175988, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Model checking and simulation of activity models using DEVS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2981823863, "categories": ["formal_verification", "computer_architecture"], "primary": "formal_verification", "confidence": 0.75, "rationale": "Hardware SAT solver with clause learning for circuit self-verification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403534925, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Regression model for dairy farm profit management; not core computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 6927776925, "categories": ["formal_verification", "algorithms"], "primary": "formal_verification", "confidence": 0.85, "rationale": "LTLf to DFA conversion tool, automata and temporal logic.", "model": "claude-sonnet-5-5"}
++{"work_id": 2904264076, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Materials science first-principles alloy study; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378551252, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.92, "rationale": "Automated proof of Petri net equivalences via SMT encoding.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405491888, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.4, "rationale": "Petri net modeling and simulation of a drone delivery system; formal modeling, performance analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 2985419616, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Philosophy/history of logic (syllogism, supposition theory), not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387635646, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.85, "rationale": "SMT string solver built on Z3, an automated reasoning tool.", "model": "claude-sonnet-5-5"}
++{"work_id": 2998652541, "categories": ["robotics"], "primary": "robotics", "confidence": 0.4, "rationale": "Sensor fusion with a Kalman filter for smartphone indoor positioning and heading estimation; localisation is the closest fit, though not strictly a robotic system.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416603322, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "Automatic structures and their algebraic characterisations, in the area of automata theory and logic.", "model": "claude-sonnet-5-5"}
++{"work_id": 3041782238, "categories": ["ai_safety", "formal_verification", "deep_learning_architectures"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Adversarial robustness of quantized networks, with SMT-based verification of robustness guarantees.", "model": "claude-sonnet-5-5"}
++{"work_id": 7172406392, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Symbolic execution and PRISM model checking for probabilistic programs; the abstract is minimal.", "model": "claude-sonnet-5-5"}
++{"work_id": 4292072122, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Max-plus polyhedra procedure for reach sets in uncertain max-plus linear systems, with a complexity improvement.", "model": "claude-sonnet-5-5"}
++{"work_id": 4210776678, "categories": ["formal_verification", "algorithms"], "primary": "formal_verification", "confidence": 0.7, "rationale": "String constraint solving algorithms, decision procedures and a benchmark framework for string solvers.", "model": "claude-sonnet-5-5"}
++{"work_id": 2919546362, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Improving PDR trajectories from pedestrian sensor data for indoor location-based services (ubiquitous computing).", "model": "claude-sonnet-5-5"}
++{"work_id": 4409052338, "categories": ["formal_verification", "algorithms"], "primary": "formal_verification", "confidence": 0.7, "rationale": "CDCL SAT solver heuristics (deep restart and a coverage metric), which is SAT solving.", "model": "claude-sonnet-5-5"}
++{"work_id": 4360995890, "categories": ["formal_verification", "software_engineering"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Modeling CPS in AADL/TCSP and transforming for formal verification of safety.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401567949, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.85, "rationale": "LTL specification analysis algorithm using model checkers.", "model": "claude-sonnet-5-5"}
++{"work_id": 7202355676, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Petri net model for logistics fleet formation; transport operations research, not core CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167214329, "categories": ["formal_verification", "software_engineering"], "primary": "formal_verification", "confidence": 0.9, "rationale": "Bounded model checker for Rust with contracts and verification of unsafe code.", "model": "claude-sonnet-5-5"}
++{"work_id": 2941019358, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Circuit relations and ZX-calculus axiomatization for quantum circuits.", "model": "claude-sonnet-5-5"}
++{"work_id": 2888263960, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Materials science on high-entropy films; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3036333943, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.88, "rationale": "Theorem proving in HOL Light to verify robotic cell injection systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 2980088699, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.4, "rationale": "Categorical semantics (differential categories, linear logic); theoretical semantics, closest to programming languages.", "model": "claude-sonnet-5-5"}
++{"work_id": 2809414278, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.7, "rationale": "Complex event processing over streams with timed automata pattern matching.", "model": "claude-sonnet-5-5"}
++{"work_id": 2969344391, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Information model for accountability requirements with consistency checking via ASP.", "model": "claude-sonnet-5-5"}
++{"work_id": 2810295409, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.55, "rationale": "Formal-logic algorithm for checking observability of DES specification languages, using automata and formal languages; AUV is only the case study.", "model": "claude-sonnet-5-5"}
++{"work_id": 2942235760, "categories": ["formal_verification", "blockchain"], "primary": "formal_verification", "confidence": 0.9, "rationale": "Deductive verification of Ethereum smart contracts with Why3.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414118222, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Constraint solving/unification modulo commutativity with similarity, with proofs of soundness and completeness; automated reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416018542, "categories": ["llm", "formal_verification"], "primary": "llm", "confidence": 0.6, "rationale": "Benchmark analysis and correction for autoformalization and theorem proving with LLM pipelines in Lean.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391021938, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control-theory interval observer design for hybrid systems; not a CS category in this ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402325914, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Safety-critical control barrier function controllers, applied to autonomous trucks and connected vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416534401, "categories": ["formal_verification", "blockchain"], "primary": "formal_verification", "confidence": 0.92, "rationale": "Model checking with SPIN and LTL of Ethereum Beacon Chain epoch processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3098615674, "categories": ["algorithms", "formal_verification"], "primary": "algorithms", "confidence": 0.5, "rationale": "Answer set programming solving efficiency, optimization rewritings and lazy-grounding; automated reasoning with some verification of transformations.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128294598, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.4, "rationale": "Petri net modelling and supervisory control of airport traffic, a discrete-event formal model.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401630734, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.4, "rationale": "V-nets formalism for discrete-event diagnosis, compared with Petri nets, timed automata and chronicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 3158571805, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.3, "rationale": "Event-driven model predictive control with hybrid system safety verification. This is mainly control theory, so the fit to the ontology is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415180303, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.4, "rationale": "Sim2real gap quantification with formal guarantees for controllers, demonstrated on a Mecanum bot and a pendulum.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400077792, "categories": ["robotics"], "primary": "robotics", "confidence": 0.35, "rationale": "Interceptor guidance and trajectory optimisation for autonomous vehicles. It is essentially control and aerospace work, so the match is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396882310, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Resource semantics for bunched implications and linear logic, the foundations of separation logic and program verification.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124134160, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Verifies opacity properties of discrete event systems using automata and diagnosers.", "model": "claude-sonnet-5-5"}
++{"work_id": 2797906474, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.3, "rationale": "Petri net model for uncertain knowledge representation. It is a formal modelling framework with a weak tie to verification.", "model": "claude-sonnet-5-5"}
++{"work_id": 2947544026, "categories": ["formal_verification", "cybersecurity"], "primary": "formal_verification", "confidence": 0.7, "rationale": "Runtime monitoring of metric temporal logic policies, applied to malware and intrusion detection on Android and autonomous vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391985801, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.55, "rationale": "New multiclock logic with assume-guarantee contracts to prove stability of layered control architectures.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391937166, "categories": ["formal_verification", "robotics"], "primary": "formal_verification", "confidence": 0.8, "rationale": "CSP-based formal modelling and verification of behavior trees used in robotics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4366551135, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.7, "rationale": "Logical framework extending LF with rational terms and coinductive definitions, a type-theoretic design contribution with decidable type checking.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406549421, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.93, "rationale": "Low-light image enhancement with a lightweight network is image restoration, which belongs to computer vision.", "model": "claude-sonnet-5-5"}
++{"work_id": 2999748008, "categories": ["computer_vision", "nlp"], "primary": "computer_vision", "confidence": 0.55, "rationale": "Pipeline of image denoising, Brahmi character recognition and translation to Sinhala; the recognition work dominates, with an NLP component.", "model": "claude-sonnet-5-5"}
++{"work_id": 3206423720, "categories": ["ai_safety", "computer_vision"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Transferable adversarial attack on image classifiers via brightness transformation; adversarial robustness evaluation on ImageNet.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403045668, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.92, "rationale": "Aircraft detection in satellite imagery with CNN classifiers is remote-sensing object detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163998453, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.93, "rationale": "Single-image dehazing network is image restoration.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213494877, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Image classification framework using residual networks and strong data augmentation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2996931328, "categories": ["deep_learning_architectures", "computer_vision"], "primary": "deep_learning_architectures", "confidence": 0.75, "rationale": "Efficient CNN architectures for inference under the Micronet Challenge, which concerns model compression and efficiency.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164124236, "categories": ["multimodal_ai", "computer_graphics", "ai_agents"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Evaluates VLMs that synthesize Blender code for 3D reconstruction and editing, using RAG and an agent workflow.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404954354, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.88, "rationale": "Thyroid nodule ultrasound diagnosis comparing CNN, ViT and hybrid models is medical image analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4298009864, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.9, "rationale": "E-commerce product recognition and segmentation with an attention-modified YOLACT.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203466192, "categories": ["generative_ai", "foundation_models", "computer_vision"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Evaluates controllable retinal image synthesis from foundation model latent spaces; centered on generation, with foundation models and medical imaging as secondary.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417350489, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.75, "rationale": "Interprets self-attention via image filtering and proposes transformer architectural modifications.", "model": "claude-sonnet-5-5"}
++{"work_id": 4366827124, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.95, "rationale": "Instance segmentation network for uterine MRI, a medical image analysis task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400230393, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.55, "rationale": "CNN-based prediction for adaptive resolution in intra video coding. Image/video processing, with no better-fitting category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155178947, "categories": ["generative_ai", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.85, "rationale": "Text-driven dance motion generation with a new dataset and a motion transformer model.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402917109, "categories": ["computer_vision", "computer_graphics"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Depth-supervised Gaussian Splatting for 3D reconstruction and novel view synthesis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381948429, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Person detection with YOLO CNN to automate classroom appliances; applied detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7112321703, "categories": ["hci"], "primary": "hci", "confidence": 0.7, "rationale": "Marker-based augmented reality mobile app for virtual shoe try-on, with a focus on user requirements and design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2886945075, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Text classification of newsgroups using CNN and GloVe embeddings.", "model": "claude-sonnet-5-5"}
++{"work_id": 2950134147, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.9, "rationale": "Gait-based age estimation with a multi-stage CNN.", "model": "claude-sonnet-5-5"}
++{"work_id": 3215388118, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "GAN for WiFi CSI activity recognition under domain change; generative model used for sensing, ubiquitous computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390361966, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "CNN with optimization-based feature selection for copy-move image forgery detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388667087, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Segmentation of choroidal vasculature in OCT volumes via thresholding; medical image analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112061614, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Weighted nearest neighbor image classification for tag recommendation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416442992, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Diffusion-based framework generating terrain models from surface models; geospatial vision application.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392442054, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Watermarking of medical images for integrity, authenticity and traceability.", "model": "claude-sonnet-5-5"}
++{"work_id": 2907431110, "categories": ["robotics", "computer_vision"], "primary": "robotics", "confidence": 0.45, "rationale": "Camera-based vehicle-to-vehicle positioning and Kalman tracking via visible light communication.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386257148, "categories": ["edge_computing", "computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "IoT wearable system for dementia patient monitoring; health informatics with IoT platform.", "model": "claude-sonnet-5-5"}
++{"work_id": 2799322609, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Monocular camera pose estimation and tracking with Mean-Shift and ArUco markers.", "model": "claude-sonnet-5-5"}
++{"work_id": 3194681712, "categories": ["hci", "deep_learning_architectures"], "primary": "hci", "confidence": 0.5, "rationale": "Radar-based gesture recognition and segmentation using CNN; interaction sensing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408017987, "categories": ["computer_vision", "generative_ai"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Image processing framework with edge detection, style transfer and compression for art images; mostly image processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167467722, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.85, "rationale": "Overview of neural network architectures (CNN, RNN, LSTM) and training techniques.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416620171, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Deep learning image classification and detection of microorganisms in Gram stain slide images (medical image analysis).", "model": "claude-sonnet-5-5"}
++{"work_id": 4393212678, "categories": ["ai_safety", "computer_vision"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Analyzes inconsistencies of saliency-map explanations in XAI for medical imaging; interpretability focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 2949391068, "categories": ["deep_learning_architectures", "computer_vision"], "primary": "deep_learning_architectures", "confidence": 0.75, "rationale": "Proposes a CNN architecture modification with learned dynamic scaling policy, evaluated on vision tasks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313830031, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.9, "rationale": "Plant leaf recognition using handcrafted LBP/HOG feature fusion and ELM classifier.", "model": "claude-sonnet-5-5"}
++{"work_id": 7152933672, "categories": ["generative_ai", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Survey of generative video trailer synthesis with diffusion, autoregressive transformers and text-to-video models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400934125, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Industrial image classification with stochastic configuration networks and multi-scale features.", "model": "claude-sonnet-5-5"}
++{"work_id": 2968866686, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.9, "rationale": "Fine-grained image recognition with an attention-based CNN model.", "model": "claude-sonnet-5-5"}
++{"work_id": 2968914950, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Medical image segmentation of psoriasis skin regions using color space analysis and k-means.", "model": "claude-sonnet-5-5"}
++{"work_id": 3090375472, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.93, "rationale": "Low-light image enhancement network; image restoration and enhancement belongs to computer vision.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388642403, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Unsupervised domain adaptation for 3D MRI classification is medical image analysis, with biomedical informatics relevance.", "model": "claude-sonnet-5-5"}
++{"work_id": 7199635077, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Passive Wi-Fi radar fall detection for assistive living, using ML on sensor signals; this is assistive and ubiquitous sensing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140831868, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.88, "rationale": "Real-time face and person detection with age and gender estimation using YOLO and MobileNet.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415102501, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Flow-based generative perceptual image compression with tokenization and entropy modelling; generative reconstruction is central, with a VLM used as an optional enhancement.", "model": "claude-sonnet-5-5"}
++{"work_id": 3116998496, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Basic gait acceleration-sensor data acquisition and correlation analysis with no clear contribution in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 2769131622, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.85, "rationale": "A CNN architecture for classifying diabetic foot ulcer images; this is medical image classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411552410, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.92, "rationale": "Cattle face recognition using metric learning on image embeddings.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387664177, "categories": ["computer_vision", "hci"], "primary": "computer_vision", "confidence": 0.6, "rationale": "AR registration via camera pose estimation optimized with SA-PSO; this is pose estimation for AR.", "model": "claude-sonnet-5-5"}
++{"work_id": 3010412574, "categories": ["computer_vision", "ml_theory"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Parameter tuning of zero-shot learning with a semantic autoencoder for image recognition.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410134551, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Business and societal book on AI and innovation; not a CS research contribution in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7168123153, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Systematic review of AI in mathematics education; education research, not CS research with a fitting category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3168036902, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Turbomachinery blade design optimized with PSO, a metaheuristic; mainly engineering design, so the fit is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 4318620721, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.55, "rationale": "Automated discovery of kinetic rate models via symbolic regression and genetic programming; equation discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 6929444221, "categories": ["scientific_computing", "autonomous_scientific_discovery"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Constitutive neural networks for automated discovery of material models with uncertainty; scientific ML and surrogate modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169848785, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.3, "rationale": "Preface to a symbolic and generative AI for science venue; no abstract, so low confidence.", "model": "claude-sonnet-5-5"}
++{"work_id": 7156328285, "categories": ["autonomous_scientific_discovery", "scientific_computing"], "primary": "autonomous_scientific_discovery", "confidence": 0.5, "rationale": "Genetic programming symbolic regression for discovering governing equations (Lorenz) under noise; equation discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414898387, "categories": ["autonomous_scientific_discovery", "generative_ai", "distributed_computing"], "primary": "autonomous_scientific_discovery", "confidence": 0.7, "rationale": "Generative AI with active learning in a distributed workflow for closed-loop discovery of molecules for carbon capture.", "model": "claude-sonnet-5-5"}
++{"work_id": 3003447096, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Data conversion and metadata structuring for a materials data repository; data management and reproducibility.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410289386, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.65, "rationale": "Physically informed neural network representation of coupled-cluster downfolding effective interactions; computational chemistry with ML.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162445105, "categories": ["autonomous_scientific_discovery", "ai_agents", "llm"], "primary": "autonomous_scientific_discovery", "confidence": 0.88, "rationale": "Agent harness of skills and tools enabling AI coding agents to run atomistic research campaigns toward autonomous AI scientists.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405426097, "categories": ["ml_theory", "autonomous_scientific_discovery"], "primary": "ml_theory", "confidence": 0.55, "rationale": "Principled Wasserstein-based dissimilarity measure for expressions with consistency guarantees, motivated by equation discovery and symbolic regression.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402242289, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Review of ML/DL models for predicting chemical carcinogenicity, a toxicology and cheminformatics application.", "model": "claude-sonnet-5-5"}
++{"work_id": 7113050848, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.6, "rationale": "Demonstrates Deep Research AI tools conducting limited autonomous research and studies.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164296032, "categories": ["deep_learning_architectures", "scientific_computing"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Topology-aware message-passing GNN with persistent homology gating, applied to MOF property prediction.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214605398, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Psychometric instrument for teachers' beliefs about AI in math teaching; education research, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399747476, "categories": [], "primary": null, "confidence": 0.6, "rationale": "General overview of AI trends, startups and Industry 4.0 without a clear specific research contribution to any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171536644, "categories": ["autonomous_scientific_discovery", "reinforcement_learning"], "primary": "autonomous_scientific_discovery", "confidence": 0.8, "rationale": "Proposes efficiency evaluation for autonomous research systems and an adaptive bandit-based search procedure.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412068170, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.55, "rationale": "Conceptual classification of AI in chemistry up to AI Researcher forming hypotheses and designing substances.", "model": "claude-sonnet-5-5"}
++{"work_id": 7113026192, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education dissertation on AI-based personalized fraction learning; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117871859, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Position paper on AI tools in mathematics education; education research, not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7139945210, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.4, "rationale": "Pure mathematics (number theory/algebraic geometry) paper, but notable for AI-assisted proofs with Gemini Deep Think and AlphaEvolve; borderline on being CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411799567, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.55, "rationale": "Opinion piece on AI aiding the research process from hypothesis generation to publishing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414528656, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Qualitative language-education study of YouTube use; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154965180, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Soft-matter mechanics of mushroom foods; materials science, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403160419, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Proves completeness of SO(3)-covariant features for molecular ML and gives efficient equivariant operations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378470143, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Deep learning algorithm to discover continuous symmetry groups and Lie algebras in data.", "model": "claude-sonnet-5-5"}
++{"work_id": 2896073529, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Announcement of an essay contest; not research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226190677, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Inverse design of metasurfaces using optimization and ML surrogate models; computational engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 7207633319, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Descriptive qualitative study of AI use in primary school science classes; education research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127595070, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Educational workshop design for scientific writing; not core computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164884815, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.45, "rationale": "AI Index report covering governance, evaluation, and AI's broad impact; the governance and evaluation focus is the closest fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155543779, "categories": ["deep_learning_architectures", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Proposes a ViT variant with a positional embedding module and classifier tail for XRD phase identification, a materials-science application.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385567506, "categories": ["scientific_computing", "autonomous_scientific_discovery"], "primary": "scientific_computing", "confidence": 0.55, "rationale": "Benchmark toolkit for inverse design algorithms in nanophotonics, built on a physics simulator and aimed at scientific design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3131458733, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Survey-based SEM study of online learning adoption; an education and social science study with only incidental neural network use.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406949642, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.55, "rationale": "AI-driven discovery of electronic materials and accelerated high-throughput discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 7121660003, "categories": [], "primary": null, "confidence": 0.4, "rationale": "AutoML benchmarking for business analytics; the contribution is business application rather than a clear CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391244979, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Self-conditioned diffusion model that generates voxel-based mechanical metamaterials.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204883563, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.85, "rationale": "Harness toolkit and benchmark integration for LLM data-science agents.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163871521, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.35, "rationale": "Policy analysis on AI and data ownership in open science; it touches AI governance but is mostly legal and policy commentary.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416200173, "categories": ["computer_graphics", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Differentiable topology optimization with a CNN surrogate for homogenized stiffness of metamaterials; computational fabrication and engineering simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389566447, "categories": ["autonomous_scientific_discovery", "reinforcement_learning"], "primary": "autonomous_scientific_discovery", "confidence": 0.8, "rationale": "Deep RL symbolic regression for automated discovery of physical laws with units constraints.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200490276, "categories": [], "primary": null, "confidence": 0.4, "rationale": "AutoML pipeline for the electric power industry is an applied ML system, not clearly in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165673775, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Engineering education pedagogy study, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7211978759, "categories": ["llm", "ai_safety"], "primary": "llm", "confidence": 0.8, "rationale": "Audits frontier LLMs for retrieval versus prediction of molecular properties; an evaluation of LLMs, including memorization.", "model": "claude-sonnet-5-5"}
++{"work_id": 7170193477, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "LIMS architecture with a PostgreSQL stack and workflow management for clinical assays; a data management system.", "model": "claude-sonnet-5-5"}
++{"work_id": 3214987860, "categories": ["deep_learning_architectures", "computational_biology"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Data augmentation library for GNN and ML property prediction on chemical structures.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416999558, "categories": ["autonomous_scientific_discovery", "ml_theory"], "primary": "autonomous_scientific_discovery", "confidence": 0.5, "rationale": "Symbolic regression discovers an interpretable constitutive equation from few samples in materials engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 7217524975, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Computer-assisted instruction for school education, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404678493, "categories": ["multimodal_ai", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Multimodal neural network surrogate for predicting nanoparticle properties, replacing DFTB-level simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4224247694, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.3, "rationale": "Physical reservoir computing in a metastructure for wave control is mostly mechanical engineering, with some learning framing. The fit to any category is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410956047, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.3, "rationale": "Time-series forecasting for PV power using EEMD, fuzzy cognitive maps and PSO. This is an energy application with a weak fit to the ontology; the closest match is the neural/learning model design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415102374, "categories": ["deep_learning_architectures", "sustainable_computing"], "primary": "deep_learning_architectures", "confidence": 0.55, "rationale": "Proposes a deterministic algebraic method for computing weights as an alternative to gradient descent, and claims energy savings and green AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011306938, "categories": ["deep_learning_architectures", "privacy_preserving_computation"], "primary": "deep_learning_architectures", "confidence": 0.55, "rationale": "GNN link prediction for master data management, including anonymization, explainability and verification.", "model": "claude-sonnet-5-5"}
++{"work_id": 2978220519, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Reservoir computing applied to generalized synchronization in chaotic systems. This is dynamical-systems analysis of a recurrent learning model.", "model": "claude-sonnet-5-5"}
++{"work_id": 7197424828, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.9, "rationale": "Training methods for quantized and binary neural networks using stochastic quantization.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312870744, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "Fast multipliers designed for hardware implementation of neural networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403331393, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Optimal power flow in electrical power systems solved with PSO and ANN. This is electrical engineering optimization; the fit to the ontology is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118589965, "categories": ["deep_learning_architectures", "computer_vision", "edge_computing"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Pruning combined with meta-learning for few-shot plant disease image classification, deployed on a Raspberry Pi edge device.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407405188, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "IMM-LSTM state fusion for maneuvering target tracking. This is signal processing with an LSTM-based model.", "model": "claude-sonnet-5-5"}
++{"work_id": 7126095475, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied ML for education dropout prediction; no ontology category's main contribution fits clearly.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391892579, "categories": ["information_retrieval", "deep_learning_architectures"], "primary": "information_retrieval", "confidence": 0.45, "rationale": "Hypergraph learning with query-guided discovery of POI strategic groups; query-driven retrieval and GNN-style modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413152908, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.8, "rationale": "Bias-variance-correlation tradeoff, a statistical learning theory topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400948005, "categories": [], "primary": null, "confidence": 0.35, "rationale": "Applied deep learning for crypto quantitative investment with sentiment factors; no category clearly fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 3042576877, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Learning rules for PCA with objective function, fixed point and stability analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416583002, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Transformer-based super-resolution of biomass maps fused with multispectral remote-sensing data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408447422, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "Deep learning automated essay scoring of English writing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4296588701, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Graph neural network construction for traffic mobility prediction with IoT data; mostly applied.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402273825, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Lyapunov-Krasovskii stability criterion via LMIs for delayed neural networks; formal mathematical analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 2797984158, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Multi-kernel semi-supervised extreme learning machine; kernel method with regularised optimisation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7141530726, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "Hybrid electrochemical model solved by orthogonal collocation plus a neural network for battery SoC estimation; a numerical-model and ML hybrid for engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148783783, "categories": ["deep_learning_architectures", "quantum_computing", "computer_vision"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Thesis on CNN, STN and LSTM variants for handwriting recognition plus a hybrid classical-quantum neural network.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416177815, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Hybrid VMD-KAN-Informer deep model for load forecasting; mainly an application-specific architecture combination.", "model": "claude-sonnet-5-5"}
++{"work_id": 3159894882, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "GPU acceleration of graph sampling for graph ML training, a hardware-aware systems contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391810135, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Piano finger coordination modeling with a BP neural network; biomechanics application, not clearly in any CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3031233905, "categories": ["ai_accelerators", "speech_audio"], "primary": "ai_accelerators", "confidence": 0.55, "rationale": "Spin-wave physical reservoir computing hardware tested on spoken digit recognition; unconventional computing hardware for AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391129611, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "PINN applied to transient gas pipeline flow and compared with finite difference method; scientific machine learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407360793, "categories": ["scientific_computing", "deep_learning_architectures"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Physics-informed GNN as a fast surrogate for AC power flow solvers.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393372557, "categories": ["computational_biology", "edge_computing"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Hyperdimensional computing for classifying fMRI brain graphs; a biomedical data method, with a low-power edge motivation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387116566, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Deep survival models for cancer risk prediction from a biobank cohort; biomedical and health informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391129094, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "BP neural network optimized with a genetic algorithm for prediction on chemical reaction data; mostly applied neural network training.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414859951, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Physical reservoir computing demonstration in a nanomechanical resonator; no abstract, appears physics rather than CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161991053, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Health informatics study comparing LSTM/Transformer and frequency models on simulated claims event sequences.", "model": "claude-sonnet-5-5"}
++{"work_id": 3195270160, "categories": ["deep_learning_architectures", "algorithms"], "primary": "deep_learning_architectures", "confidence": 0.55, "rationale": "Applies GNNs to the Steiner tree combinatorial problem and compares against 2-approximation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386990763, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.85, "rationale": "Knowledge-graph and deep-learning recommender for tourist attractions evaluated with ranking metrics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220870911, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Overview of AI history and its role in education, with a mismatched title; not clearly research in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7109160285, "categories": ["ml_theory", "ai_agents"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Variational principle deriving softmax gating as an energy-based dynamical system with an RNN instantiation, for behavioral composition.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378192806, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Cluster-based LSTM time series forecasting for dengue; applied public health forecasting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410775015, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.35, "rationale": "CNN for breast cancer prediction on tabular Wisconsin data with hyperparameter tuning; applied.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407130511, "categories": ["deep_learning_architectures", "ai_safety"], "primary": "deep_learning_architectures", "confidence": 0.8, "rationale": "Thesis on graph neural differential equation networks for representation learning and adversarial robustness.", "model": "claude-sonnet-5-5"}
++{"work_id": 3047152076, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Theoretical physics-style treatment of learning dynamics of neural networks as a model of the universe; loosely ML theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4380551192, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.8, "rationale": "Proves results about parameter-space symmetries and functional dimension of ReLU networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405619363, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Cultural analysis of generative AI music systems and their impact; weak CS research but concerns generative AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 7149698496, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Graph-based multi-channel fusion network with Graph Transformer for fault diagnosis under imbalance.", "model": "claude-sonnet-5-5"}
++{"work_id": 2802497406, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.85, "rationale": "CNN pruning and compression framework.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408222477, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.92, "rationale": "Adaptive learning rate scheduler for federated learning using PID control.", "model": "claude-sonnet-5-5"}
++{"work_id": 4283646524, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Deep-learning classification of esophageal lesions in endoscopic images, medical image analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389157323, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Wind speed forecasting with a TCN-GRU model and denoising; an application of neural architectures.", "model": "claude-sonnet-5-5"}
++{"work_id": 2907877223, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.45, "rationale": "Multi-reservoir echo state network (recurrent architecture) for fuel cell remaining useful life prediction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414684252, "categories": ["scientific_computing", "deep_learning_architectures"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Hypernetwork surrogate for forecasting parametric dynamical systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3183119475, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Theoretical analysis of co-evolving opinion and signed-network dynamics; the closest fit is algorithms/combinatorics, but the paper is only loosely CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2898975570, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.95, "rationale": "Analysis of Quickselect variants with multiple pivots and cache efficiency.", "model": "claude-sonnet-5-5"}
++{"work_id": 2895538521, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Spectral graph theory in pure discrete mathematics, not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401726474, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Recovery guarantees for sparse tensors via OMP and mutual incoherence; this is a theoretical sparse-recovery and information-theory result.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400170167, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Search algorithm for efficient retrieval over large datasets, using deep learning and distributed computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385209635, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Automatic QUBO formulations for combinatorial optimization on Ising machines.", "model": "claude-sonnet-5-5"}
++{"work_id": 7115725860, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.9, "rationale": "Online algorithms with competitive analysis and metaheuristics for dynamic routing and optimization problems.", "model": "claude-sonnet-5-5"}
++{"work_id": 2998687774, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.9, "rationale": "Book on theoretical algorithms for big data, including streaming.", "model": "claude-sonnet-5-5"}
++{"work_id": 4205264571, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.92, "rationale": "Single-sample prophet inequalities and online selection, with algorithmic game theory and mechanism design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313546928, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Evolutionary algorithm for dynamic multiobjective optimization, using neural-network prediction, with a multirobot scheduling application.", "model": "claude-sonnet-5-5"}
++{"work_id": 3040673611, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Memetic swarm optimizer (metaheuristic) for a combinatorial network optimization problem; fits combinatorial optimisation and metaheuristics.", "model": "claude-sonnet-5-5"}
++{"work_id": 3162282920, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Applies the Harris Hawks metaheuristic to solve Hyper Sudoku; a metaheuristic combinatorial optimisation study.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415799037, "categories": ["algorithms", "scientific_computing"], "primary": "algorithms", "confidence": 0.7, "rationale": "Provably sublinear-time local solvers for graph diffusion equations; graph algorithm design with a numerical-solver flavour.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415332021, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.8, "rationale": "Query-efficient graph reconstruction and seriation algorithm with provable guarantees.", "model": "claude-sonnet-5-5"}
++{"work_id": 3197190047, "categories": ["algorithms", "computer_networks"], "primary": "algorithms", "confidence": 0.55, "rationale": "Multiobjective evolutionary algorithm that optimizes robustness and transmission capacity of networks; the algorithm is the main contribution, with a network-design application.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163898549, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Addition-multiplication chains and bounds on their lengths; combinatorics and number theory underlying computation of powers.", "model": "claude-sonnet-5-5"}
++{"work_id": 2890999007, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Quantum-inspired local search combined with binary PSO; metaheuristic optimisation, not quantum computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7218127097, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Graph theory result on distance domination numbers of corona graphs; combinatorics that underpins algorithms.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400392879, "categories": [], "primary": null, "confidence": 0.4, "rationale": "No abstract; a popular-science style title, so there is not enough information to assign a category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3151742782, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.9, "rationale": "Dissertation on adaptive approximation algorithms for stochastic submodular ranking, routing and decision trees.", "model": "claude-sonnet-5-5"}
++{"work_id": 2911086832, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "A novel metaheuristic optimizer applied to PID tuning; the contribution is a heuristic optimization algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 3200096258, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.35, "rationale": "Computational algebraic geometry of tensor varieties, published in the Journal of Symbolic Computation; this is the nearest fit, though it is mostly pure mathematics.", "model": "claude-sonnet-5-5"}
++{"work_id": 3136531308, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.85, "rationale": "Computing the continuous mean distance of a weighted graph using graph and computational-geometry algorithms, with time bounds.", "model": "claude-sonnet-5-5"}
++{"work_id": 7130513340, "categories": ["computational_complexity", "algorithms"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "Proves NP-completeness for k-rainbow total domination and gives graph bounds and a counterexample to a conjecture.", "model": "claude-sonnet-5-5"}
++{"work_id": 6947983997, "categories": ["algorithms", "ml_theory"], "primary": "algorithms", "confidence": 0.6, "rationale": "Online learning of bidding strategies in karma auctions, with equilibrium analysis; combines algorithmic game theory and online learning theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392544619, "categories": [], "primary": null, "confidence": 0.6, "rationale": "A description of a matrix operation, which is mathematical and not clearly computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3180083090, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "A new evolutionary algorithm (an EDA, UMDA variant) for smart grid optimization problems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4254111132, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.55, "rationale": "A tutorial on representations in evolutionary algorithms, a metaheuristics topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 2981962366, "categories": [], "primary": null, "confidence": 0.6, "rationale": "A computing-education study of student strategies, which fits no category in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414079667, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "An engineering design optimization using an improved simulated annealing algorithm with a Kriging surrogate model.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396597290, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.4, "rationale": "Adaptive dynamic programming for differential game guidance; aerospace control, closest to RL/ADP.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409058110, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Review of RUL prediction for rotating machinery; mechanical engineering/instrumentation, not CS-core.", "model": "claude-sonnet-5-5"}
++{"work_id": 3044771921, "categories": [], "primary": null, "confidence": 0.5, "rationale": "CS education study on teaching contexts; no ontology category fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 3087997050, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.7, "rationale": "Comparative study of metaheuristic optimization algorithms.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203700339, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.92, "rationale": "W[1]-hardness and ETH lower bounds for norm maximization.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410943262, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Applies multi-objective metaheuristics to a chemical reactor optimization; mainly chemical engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 2758153130, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Monte Carlo method for bounds on stochastic dynamic programming; numerical/financial math.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404370937, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.65, "rationale": "Differential evolution metaheuristic for constrained optimization with IoT layout application.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161020069, "categories": ["algorithms", "computational_complexity"], "primary": "algorithms", "confidence": 0.7, "rationale": "Social choice efficiency analysis with NP-completeness, APX-hardness and FPT algorithms.", "model": "claude-sonnet-5-5"}
++{"work_id": 2978556928, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.85, "rationale": "Metaheuristic (CRO) for the maximum independent set problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 2974268229, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.9, "rationale": "Online local routing in sparse planar geometric graphs; geometric graph algorithm design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035652685, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Graph labeling results (combinatorics of graphs); closest fit is the combinatorics underpinning algorithms, though it is mostly pure math.", "model": "claude-sonnet-5-5"}
++{"work_id": 3123947211, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Smoothing function and Jacobian consistency for complementarity problems; numerical optimization methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4320041016, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.45, "rationale": "Graph domination number bounds, combinatorics of graphs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415250135, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Antenna array optimization with CRLBs is signal processing/electromagnetics, not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7024143063, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.93, "rationale": "Quasi-polynomial time algorithms for Independent Set in hereditary graph classes.", "model": "claude-sonnet-5-5"}
++{"work_id": 2982343211, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.7, "rationale": "Exact algorithm via MILP for arrow placement in graph drawing; combinatorial optimization.", "model": "claude-sonnet-5-5"}
++{"work_id": 3115336282, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Interior-point algorithm for linear optimization with iteration complexity analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3166282595, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Code and data on estimating metaheuristic ability to guide heuristics; combinatorial optimization metaheuristics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413144369, "categories": ["algorithms", "computer_architecture"], "primary": "algorithms", "confidence": 0.5, "rationale": "Compressed QUBO formulation for TSP on Ising machines; combinatorial optimization with a hardware-oriented angle.", "model": "claude-sonnet-5-5"}
++{"work_id": 7115594566, "categories": ["llm", "nlp", "cybersecurity"], "primary": "llm", "confidence": 0.6, "rationale": "Prompt-engineering framework using GPT to build knowledge graphs/ontologies, with a cybersecurity case study.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399491257, "categories": ["llm", "hci"], "primary": "llm", "confidence": 0.5, "rationale": "Custom GPT chatbot as an educational resource, evaluated through expert interviews and student surveys.", "model": "claude-sonnet-5-5"}
++{"work_id": 7113152375, "categories": ["llm", "information_retrieval"], "primary": "llm", "confidence": 0.6, "rationale": "RAG-based university chatbot built with an LLM and a hybrid knowledge base.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402670518, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.6, "rationale": "Synthetic annotated text generation from knowledge graphs with LLMs for information extraction.", "model": "claude-sonnet-5-5"}
++{"work_id": 7149228871, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.5, "rationale": "Dataset for distinguishing human-written from LLM-generated metaphors via annotation and classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416889670, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Non-CS essay on simplified math methods and education; no clear computer-science contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221549073, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.85, "rationale": "Mechanistic analysis of refusal and jailbreak pathways in aligned LLMs.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161067881, "categories": ["llm", "information_retrieval"], "primary": "llm", "confidence": 0.7, "rationale": "Benchmark and system for LLM memory retrieval precision.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131836878, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.8, "rationale": "Prompting framework improving LLM reasoning on math word problems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387799763, "categories": ["llm", "ai_safety", "nlp"], "primary": "llm", "confidence": 0.75, "rationale": "Evaluates the faithfulness of ChatGPT self-explanations against traditional explanation methods on sentiment analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409475173, "categories": ["llm", "information_retrieval"], "primary": "llm", "confidence": 0.8, "rationale": "RAG-based LLM educational chatbot evaluated with retrieval and faithfulness metrics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394729406, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.75, "rationale": "BERT-based keyword search for movies using reviews and rankings; a retrieval task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416454327, "categories": ["llm", "ai_safety"], "primary": "llm", "confidence": 0.8, "rationale": "Benchmark and metrics for evaluating cultural understanding and alignment in LLMs.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138061291, "categories": ["llm", "software_engineering"], "primary": "llm", "confidence": 0.7, "rationale": "Prompting strategies with execution feedback for LLM code generation in a shared task; Bangla input.", "model": "claude-sonnet-5-5"}
++{"work_id": 3109984837, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Palmprint biometric identification via geometric analysis of palm images.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404852650, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "BERT and CNN hybrid models for Chinese scientific paper text classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165886470, "categories": ["deep_learning_architectures", "llm"], "primary": "deep_learning_architectures", "confidence": 0.75, "rationale": "Analysis of looped transformer recurrence, normalization and readout supervision; architectural focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128818302, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Fine-tuned T5 abstractive summarization of social media text with an evaluation framework.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214900323, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Survey of student LLM use in language learning; education research, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414501685, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.6, "rationale": "Automated review labeling combining keyword detection and LLM prompting or fine-tuning; text classification framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416245062, "categories": ["llm", "ai_accelerators", "ml_theory"], "primary": "llm", "confidence": 0.8, "rationale": "Online draft model selection for speculative decoding in LLM inference, with regret guarantees and system-efficient implementations.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117738325, "categories": ["autonomous_scientific_discovery", "llm", "reinforcement_learning"], "primary": "autonomous_scientific_discovery", "confidence": 0.8, "rationale": "Trains AI co-scientist models to generate research plans via RL with rubric self-grading.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391976463, "categories": ["nlp", "llm"], "primary": "nlp", "confidence": 0.7, "rationale": "Human evaluation of ChatGPT's machine translation quality across domains.", "model": "claude-sonnet-5-5"}
++{"work_id": 7110969657, "categories": ["llm", "nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Evaluates ChatGPT on Romanian verb conjugation, a computational linguistics and LLM evaluation task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414790785, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.8, "rationale": "Architecture and study of autonomous LLM agent behavior without external tasks.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214894185, "categories": ["llm", "information_retrieval"], "primary": "llm", "confidence": 0.7, "rationale": "Enhances oncology LLMs with a retrieved tree-relational structure, improving over RAG and GraphRAG.", "model": "claude-sonnet-5-5"}
++{"work_id": 7153669768, "categories": ["llm", "reinforcement_learning", "ai_safety"], "primary": "llm", "confidence": 0.65, "rationale": "Self-training of generative reward models for LLM alignment using consistency-aware rewards.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155055855, "categories": ["llm", "deep_learning_architectures"], "primary": "llm", "confidence": 0.7, "rationale": "Knowledge distillation between LLMs with different tokenizers, analyzing cross-model attention and proposing a GAN-based method.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166569973, "categories": ["multimodal_ai", "llm", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Multimodal LLM/VLM/audio framework with multimodal RAG for customer service and incident management.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171730128, "categories": ["information_retrieval", "database_systems"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Proposes human-preference-aligned evaluation of tabular embeddings for similarity search.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128492423, "categories": ["llm", "information_retrieval"], "primary": "llm", "confidence": 0.7, "rationale": "RAG-based chatbot using Mistral LLM and vector retrieval for curriculum-aligned question answering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399996928, "categories": ["nlp"], "primary": "nlp", "confidence": 0.7, "rationale": "Fine-tuning BERT for consumer sentiment and behavior analysis is a classical text classification task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387819913, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Qualitative study of pre-service teachers' use of ChatGPT in education, a user study of human-AI use.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124885328, "categories": ["llm"], "primary": "llm", "confidence": 0.92, "rationale": "Survey of domain-specific LLMs covering continual pre-training and prompt engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 3185341429, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.88, "rationale": "Survey of prompting methods for pre-trained language models, covering few-shot and zero-shot learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117486691, "categories": ["llm", "ai_agents"], "primary": "ai_agents", "confidence": 0.75, "rationale": "Overview of LLM-based agentic AI and multi-agent collaboration, with a comparison of fine-tuning and RAG.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417258277, "categories": ["llm", "ml_theory"], "primary": "llm", "confidence": 0.8, "rationale": "Controlled pre-training study of memorization versus generalization and capacity in transformer language models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403025669, "categories": ["llm"], "primary": "llm", "confidence": 0.8, "rationale": "Chapter on human evaluation methods for testing LLMs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401044042, "categories": ["multimodal_ai", "llm", "speech_audio"], "primary": "multimodal_ai", "confidence": 0.88, "rationale": "Aligns a music audio encoder with a frozen LLM for music captioning and question answering, a multimodal audio-text task.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165003687, "categories": ["llm", "nlp"], "primary": "llm", "confidence": 0.8, "rationale": "Survey of LLM-powered question answering pipelines organized by a query-processing taxonomy.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416251417, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.5, "rationale": "LLM-integrated decision framework with ReAct-style prompting and a bandit patrol-allocation algorithm for wildlife trade policy.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138971369, "categories": ["llm"], "primary": "llm", "confidence": 0.6, "rationale": "Case study of prompting LLMs to author educational tasks, with expert evaluation of the output.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125957422, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Multi-agent architecture that integrates LLMs for natural-language interaction with smart meter data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414992583, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Educational systematic review of AI tools in K-12 language teaching, grounded in sociocultural theory; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387935330, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Legal and educational discussion of AI in inclusive education, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391421531, "categories": ["llm"], "primary": "llm", "confidence": 0.92, "rationale": "Introduces a family of LLMs pre-trained and aligned for writing, with RAG and tool use support.", "model": "claude-sonnet-5-5"}
++{"work_id": 4395022812, "categories": ["llm"], "primary": "llm", "confidence": 0.88, "rationale": "Proposes a stronger random baseline for evaluating in-context learning of language models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414463094, "categories": ["nlp"], "primary": "nlp", "confidence": 0.8, "rationale": "Arabic dialect classification comparing RNNs, transformers and LLMs; the task framing dominates.", "model": "claude-sonnet-5-5"}
++{"work_id": 7199029551, "categories": ["llm"], "primary": "llm", "confidence": 0.4, "rationale": "Title concerns linguistic representations in LLMs, but there is no abstract; likely an interpretability or philosophy-of-language piece.", "model": "claude-sonnet-5-5"}
++{"work_id": 2904234310, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Representation learning to build sentiment dictionaries from customer reviews.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140096463, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Network pharmacology, docking and itemset mining on EMRs for TCM prescriptions; biomedical informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312852711, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Interpretable rule-based prediction model on EMR data for ICU stay; biomedical/health informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167456159, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Molecular docking and in silico drug candidate analysis against bacterial targets.", "model": "claude-sonnet-5-5"}
++{"work_id": 4296776208, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.88, "rationale": "Local ancestry inference method on genomic haplotype data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386760409, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Network pharmacology and docking for a herbal mixture; computational drug discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402637119, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Radiomics and deep learning for classifying breast masses in mammography/tomosynthesis images.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410249860, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "QSAR, docking and text mining workflow for phytochemical virtual libraries, with an educational use.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164761540, "categories": ["foundation_models", "computational_biology", "multimodal_ai"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Pathology foundation model aligning histopathology with DNA methylation; multimodal and biomedical.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394782468, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Digital twin simulation and ILP optimization of neoantigen vaccine composition.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400594688, "categories": ["computational_biology", "ml_theory"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Bayesian multi-layer network model for community connectivity in spatial transcriptomics data.", "model": "claude-sonnet-5-5"}
++{"work_id": 3127450329, "categories": ["hci"], "primary": "hci", "confidence": 0.45, "rationale": "Empirical study of clinician EHR proficiency after training; human factors of an interactive system, borderline CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3081811896, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.45, "rationale": "Review of computer-aided synthesis planning using AI for organic chemistry; closest to molecular modelling and drug discovery methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4308207210, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Structure-based virtual screening for drug discovery with biological assays.", "model": "claude-sonnet-5-5"}
++{"work_id": 3005160664, "categories": ["computational_biology", "computer_graphics"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Geostatistical visualization of tumor ecology with an agent-based model; bioinformatics tool.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391930238, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Review of ML and deep learning for lung nodule detection in CT and X-ray images.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389577171, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Continual learning of deep models for cuffless blood pressure estimation from PPG/ECG; biomedical informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389302773, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Free energy perturbation binding free energy calculation for structure-based drug design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411492298, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Software tool for mass spectrometry data quality checking; cheminformatics and bioinformatics tooling.", "model": "claude-sonnet-5-5"}
++{"work_id": 3163978416, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Data model mapping between SDTM and FHIR; data integration for health informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411977602, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Computational drug repurposing using docking and molecular dynamics simulations.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135254915, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "In silico molecular docking of plant phenolic compounds against enzymes; computational drug discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411893160, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Automated segmentation and morphological analysis of dendritic spine microscopy images via skeletonization.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162037404, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.75, "rationale": "Inductive gene-disease association prediction using phenotype ontology graph embeddings; bioinformatics method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388076232, "categories": ["computational_biology", "deep_learning_architectures"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Graph neural method for lncRNA-disease association prediction; biomedical application.", "model": "claude-sonnet-5-5"}
++{"work_id": 3153063642, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Computational characterization of ligand binding for drug design on a receptor.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221541424, "categories": ["computational_biology", "generative_ai"], "primary": "computational_biology", "confidence": 0.75, "rationale": "Computational drug discovery with docking, ADMET prediction and retrosynthesis, using generative analogue design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3177565309, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.8, "rationale": "Docking and molecular dynamics of designed protease inhibitors; computational drug discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4318046762, "categories": ["hci", "multimodal_ai"], "primary": "hci", "confidence": 0.5, "rationale": "Annotation tool and dataset for human-AI collaborative pathology diagnosis with captioning and classification; trust study.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221738702, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Systems biology network and biomedical data mining for cross-species drug target discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4302425986, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.65, "rationale": "Network pharmacology bioinformatics analysis of coix seed targets in colorectal cancer.", "model": "claude-sonnet-5-5"}
++{"work_id": 3174432239, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Curated database of reconstructed patient-level trial data; data resource, borderline CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160950788, "categories": ["deep_learning_architectures", "computational_biology", "scientific_computing"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Introduces a GNN (QT-Net) and an evaluation protocol for atomic property prediction in chemistry.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405278807, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Conference proceedings blurb on chemical activity prediction; no CS content evident.", "model": "claude-sonnet-5-5"}
++{"work_id": 4220748066, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Bioinformatics and network pharmacology analysis of a drug's mechanism.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411249474, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Virtual screening, docking and MD for antiviral drug discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 2783094078, "categories": ["computational_biology", "computer_networks"], "primary": "computational_biology", "confidence": 0.6, "rationale": "DNA data storage and retrieval via bacterial nanonetworks; molecular computing/communication.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135069038, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Survey of healthcare workers' perceptions of digital pathology and AI technology adoption.", "model": "claude-sonnet-5-5"}
++{"work_id": 3173374233, "categories": ["generative_ai", "computational_biology"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Autoregressive graph generative model with Metropolis sampling for molecule optimization.", "model": "claude-sonnet-5-5"}
++{"work_id": 2985643449, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Pharmacophore modelling for inhibitor identification in computational drug discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 7211895300, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "In silico docking, MM/GBSA and ADMET prediction for gyrase inhibitors.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402911702, "categories": ["ml_theory", "computational_biology"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Statistical method for sparse multivariate additive regression with joint covariance estimation; applied to proteomics data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4295929866, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Deep learning segmentation of vessels and tissue in endoscopic images.", "model": "claude-sonnet-5-5"}
++{"work_id": 3159313223, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.9, "rationale": "In silico molecular docking and ADME/T analysis for drug discovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378771166, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.88, "rationale": "Drug repurposing using molecular docking and ML QSAR regression.", "model": "claude-sonnet-5-5"}
++{"work_id": 3110762405, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.85, "rationale": "Deep neural networks predicting protein hydration thermodynamics for molecular modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412915491, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Clinical decision support system and data analysis for COVID-19; biomedical informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393899952, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.5, "rationale": "Image-processing algorithm for quantifying mitochondria in pathology images.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367393431, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.9, "rationale": "Ensemble learning to predict miRNA-disease associations; bioinformatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 3020560080, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Review of AI for breast medical imaging diagnosis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3159857369, "categories": ["ml_theory", "computational_biology"], "primary": "ml_theory", "confidence": 0.55, "rationale": "Review of regularization techniques for robust modelling in life science data.", "model": "claude-sonnet-5-5"}
++{"work_id": 2786392319, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Vague essay on AI, robots and climate; no clear CS research contribution fitting a category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3093665008, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.35, "rationale": "Literature review on uncertainty management in AI-enabled decision making; weak fit to autonomous decision systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387951271, "categories": [], "primary": null, "confidence": 0.6, "rationale": "HVAC building simulation with TRNSYS; engineering, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386212618, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.85, "rationale": "Estimates carbon emissions of website data transfer and proposes green web practices.", "model": "claude-sonnet-5-5"}
++{"work_id": 7168018191, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Media and cultural studies literature review, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410538484, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Multi-criteria decision analysis of energy sources; energy policy, not CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313647803, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Experimental study of how online review cues affect users' purchase intention; consumer behaviour, loosely HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404238308, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Review of solar power forecasting methods; energy domain application of ML, no CS category clearly fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412066060, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.92, "rationale": "Testbed measuring energy and carbon footprint of data-centre throughput computing, including AI inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403003621, "categories": ["hci", "sustainable_computing"], "primary": "hci", "confidence": 0.85, "rationale": "Focus group user study on IoT repairability and software obsolescence, with sustainability motivation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3154667522, "categories": ["sustainable_computing", "cloud_computing"], "primary": "sustainable_computing", "confidence": 0.5, "rationale": "Green IT and a cloud LMS model for a sustainable learning environment; weak CS research but touches sustainable and cloud computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215545478, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Community service teacher training on e-learning; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3111008797, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.35, "rationale": "LSTM occupancy prediction for HVAC energy saving; applied time-series ML, with building energy as the application and no better-fitting category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409494225, "categories": ["deep_learning_architectures", "ai_safety", "ai_accelerators"], "primary": "deep_learning_architectures", "confidence": 0.7, "rationale": "Neural architecture search for adversarially robust approximate DNNs on systolic-array hardware accelerators.", "model": "claude-sonnet-5-5"}
++{"work_id": 4308155833, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Remote sensing CO2 reconstruction for earth observation; domain geoscience, not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7034426089, "categories": ["software_engineering", "sustainable_computing", "formal_verification"], "primary": "software_engineering", "confidence": 0.65, "rationale": "Model-based (probabilistic LTS) analysis of software energy consumption for developers.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400064539, "categories": ["sustainable_computing", "computer_networks", "deep_learning_architectures"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Energy-aware transformer model selection for green semantic communication; focus on energy footprint.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387983484, "categories": ["software_engineering", "sustainable_computing"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Empirical study of energy consumption evolution of Kotlin Android apps across releases.", "model": "claude-sonnet-5-5"}
++{"work_id": 3080194848, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Agricultural remote sensing of soil respiration with regression; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406505865, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Organizational readiness survey for digital transformation in a university; information systems or management study, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7141737583, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Business and management book on AI strategy, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 6925471442, "categories": ["sustainable_computing", "software_engineering"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Course design on Green Coding and energy-efficient software; the topic is green software.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135017252, "categories": ["reinforcement_learning", "llm", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "LLM-generated rewards embedded in multi-agent RL for low-carbon energy scheduling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408774499, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Econometric ARDL study of emissions and economy; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4309156398, "categories": ["sustainable_computing", "cloud_computing"], "primary": "sustainable_computing", "confidence": 0.65, "rationale": "Server power estimation and management system in data centers to reduce power consumption.", "model": "claude-sonnet-5-5"}
++{"work_id": 2996084711, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.55, "rationale": "Review of smartphone power consumption, management and batteries.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220836233, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Education management library study, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409426101, "categories": ["sustainable_computing", "cloud_computing"], "primary": "sustainable_computing", "confidence": 0.75, "rationale": "Game-theoretic energy and workload management of geo-distributed data centers with renewables and emissions.", "model": "claude-sonnet-5-5"}
++{"work_id": 3001058942, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Library science introduction on technical reports, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2948997629, "categories": ["hci", "sustainable_computing"], "primary": "hci", "confidence": 0.5, "rationale": "Critical analysis of a smart city project and its social and sustainability dimensions in ICT4S.", "model": "claude-sonnet-5-5"}
++{"work_id": 3025387569, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.5, "rationale": "Essay on digital ecology and the place of digital technology in a socio-ecological transition; touches on computing's footprint.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416214949, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Power-system demand response and carbon trading for an industrial load; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386108360, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.92, "rationale": "Renewable energy sizing and cost for data centres, directly about the carbon footprint of computing infrastructure.", "model": "claude-sonnet-5-5"}
++{"work_id": 2905015752, "categories": ["sustainable_computing", "edge_computing"], "primary": "sustainable_computing", "confidence": 0.45, "rationale": "Measures performance and power consumption of CPU vs GPU image processing on smartphones; energy-efficiency evaluation of on-device computation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2950028967, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Demand-side carbon footprint management in power systems; not about computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392157044, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Econometric study of ICT, gender inclusion and CO2 emissions; economics, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4207059814, "categories": ["sustainable_computing", "software_engineering"], "primary": "sustainable_computing", "confidence": 0.75, "rationale": "Tool for estimating the energy consumption of Android apps; green software engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415684933, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Economics/finance policy study using DID; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990816793, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.85, "rationale": "Survey of e-waste from ICT devices, covering material footprint and recycling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4323665427, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Urban green space and building carbon emissions; environmental science, not computing systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399765326, "categories": [], "primary": null, "confidence": 0.85, "rationale": "PV panel yield simulation under climate factors is energy engineering, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124145190, "categories": ["sustainable_computing", "ai_accelerators"], "primary": "sustainable_computing", "confidence": 0.88, "rationale": "Energy efficiency metric for AI servers running LLM workloads on NPUs; Green AI measurement.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406559909, "categories": [], "primary": null, "confidence": 0.75, "rationale": "RBF neural network MPPT controller for PV systems is electrical engineering control, not core CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117113370, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Ecological habitat suitability assessment for dolphins; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390978249, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Provincial carbon emission forecasting with LEAP; energy policy, not computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3036320021, "categories": ["sustainable_computing", "cloud_computing", "algorithms"], "primary": "sustainable_computing", "confidence": 0.75, "rationale": "Energy-aware metaheuristic green scheduling in virtual cloud with hardware energy regulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4361987249, "categories": [], "primary": null, "confidence": 0.7, "rationale": "ANFIS-GA prediction of PV output is an energy application with standard methods; not CS research focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399284355, "categories": [], "primary": null, "confidence": 0.65, "rationale": "Forecasting agricultural carbon emissions with LSTM; environmental application, not computing footprint.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390621478, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.6, "rationale": "Awareness challenge on environmental impacts of digital technologies; education about computing footprint.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990724040, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.9, "rationale": "Reduces data-centre power losses via UPS consolidation; data-centre energy efficiency.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163329089, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational survey study of an LMS in a school; not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367163963, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Shot optimization for training quantum machine learning models, which is quantum software and QML.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225771907, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Generation, transmission and detection of space-encoded quantum states for quantum communications.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415252535, "categories": ["quantum_computing", "federated_learning"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Distributed quantum ML with knowledge-distillation aggregation across heterogeneous clients, a federated-learning setup using variational quantum circuits.", "model": "claude-sonnet-5-5"}
++{"work_id": 7116076303, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Quantum radiometric calibration of photodiodes is quantum optics and metrology, not a computational framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3206680160, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Quantum key distribution protocol with thermal sources, which falls under quantum communication.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417308968, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.92, "rationale": "Quantum circuit partitioning and circuit cutting with hybrid execution on NISQ devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131679394, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.5, "rationale": "Superconducting parametric amplifiers for qubit readout; hardware-focused quantum technology, borderline for scope.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412565510, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.75, "rationale": "Scalable chip-based ion traps as a qubit hardware platform for quantum information processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2896762452, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.45, "rationale": "Motional heating in ion-trap quantum computers studied by DFT; mostly physics, framed by quantum computing hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125762882, "categories": ["quantum_computing", "computer_networks"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Experimental QKD over deployed multicore fibre; quantum communication falls under quantum_computing, with a networking framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2964632298, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Reversible logic circuit design with quantum cost metrics; digital hardware design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4296884690, "categories": ["quantum_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.65, "rationale": "Laser damage attack on QKD hardware and a defence against it; implementation security of quantum communication systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415968028, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.4, "rationale": "Quantum metrology under noise; information-processing framing is marginal but it is quantum information.", "model": "claude-sonnet-5-5"}
++{"work_id": 2902876714, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.55, "rationale": "Trapped-ion coherent control and cooling, a qubit hardware platform for quantum computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2902785919, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Hybrid NISQ algorithm for QFT simulation, used to benchmark quantum hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163287303, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.88, "rationale": "Implementation of a quantum game on NISQ hardware with noise analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 2783787559, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.3, "rationale": "Speculative analogies between spin systems and quantum computation for molecular networks; weak fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161271332, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Entanglement detection with multi-copy circuits, validated on IBM quantum processors.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112102734, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Technical perspective on quantum error handling to reduce the cost of quantum computation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200635042, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.5, "rationale": "Entangled spin states for quantum metrology under decoherence; quantum information framing, but mostly atomic physics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312768793, "categories": ["quantum_computing", "computer_networks", "cybersecurity"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Review of QKD for visible light communication security at the physical layer.", "model": "claude-sonnet-5-5"}
++{"work_id": 2964298217, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Quantum optics interferometric metrology; physics without computational framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411232250, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.65, "rationale": "Unambiguous state discrimination for quantum communication systems under noise.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385967641, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Mean-field analysis of IBM quantum utility experiment on many-body dynamics simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411225084, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.45, "rationale": "Superconducting resonators with high quality factors as qubit hardware platform component.", "model": "claude-sonnet-5-5"}
++{"work_id": 3161468414, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Resource-efficient VQE encoding algorithm for bosonic simulations on NISQ devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400861003, "categories": ["quantum_computing", "ai_accelerators"], "primary": "quantum_computing", "confidence": 0.55, "rationale": "GPU-accelerated tensor network contraction for classical simulation of quantum circuits.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417073300, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Software for evaluating cloud quantum computers' qubits via quantum random number generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416775576, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.75, "rationale": "Entanglement witnesses for detecting Bell states, a quantum information topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035347578, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Hybrid quantum-classical reservoir computing framework for temporal ML tasks; quantum machine learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 3137888032, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.55, "rationale": "Quantum reference frames and subsystem decompositions underlying entanglement and quantum information tasks; borderline physics-focused.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391889616, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Twin-field QKD protocol with security proof; quantum key distribution is listed under quantum computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148276906, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.92, "rationale": "Hybrid variational/adiabatic quantum state preparation algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 2995678717, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Noise protection of qudit gates via continuous dynamical decoupling; quantum control and error suppression.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387681983, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.92, "rationale": "Bayesian optimization of QAOA parameters on NISQ devices; variational quantum algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387559660, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Hybrid classical-quantum state preparation for loading probability distributions, tested on IBM hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400297779, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.88, "rationale": "Quantum machine learning and the role of entangled data, with a no-free-lunch theorem.", "model": "claude-sonnet-5-5"}
++{"work_id": 2755062554, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Experimental quantum optics metrology of single-photon wave functions; physics rather than computational framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404573991, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Entanglement robustness in noisy quantum networks linked to graph connectivity; quantum information and networking.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402268320, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Noise-robust topological features of entangled states; quantum information framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410302895, "categories": ["quantum_computing", "computational_complexity", "cybersecurity"], "primary": "quantum_computing", "confidence": 0.8, "rationale": "Quantum learning algorithms with query lower bounds and an attack on a quantum PUF authentication protocol.", "model": "claude-sonnet-5-5"}
++{"work_id": 4221156287, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.92, "rationale": "Quantum error correction with the 3D surface code, non-Clifford errors and magic state preparation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4297796489, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.5, "rationale": "Photon-pair high-dimensional entanglement generation; photonic quantum information source, borderline.", "model": "claude-sonnet-5-5"}
++{"work_id": 7110562677, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Tensor network quantum circuit simulation, VQE implementation and QUBO application.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389163069, "categories": ["quantum_computing", "applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.75, "rationale": "Semi-quantum random number generation protocol with a security proof.", "model": "claude-sonnet-5-5"}
++{"work_id": 2996620206, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Photonic cluster state generation for one-way quantum computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393118756, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.85, "rationale": "Variational and NISQ algorithms run on IBM Q to probe the quantum geometric tensor.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401437211, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.6, "rationale": "Simulation of decoherence in superconducting qubits, a training/visualization program.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127900880, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Quantum algorithm for spectral functions and state preparation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390533496, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Feature stores manage and serve data for ML pipelines, which is a data-management problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162347486, "categories": ["operating_systems", "computer_architecture"], "primary": "operating_systems", "confidence": 0.55, "rationale": "Address mapping in the SSD flash translation layer is a storage and file-system topic, with some overlap with storage hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394871762, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.9, "rationale": "Tooling to build OS compatibility layers and measure the OS features applications need.", "model": "claude-sonnet-5-5"}
++{"work_id": 3023238978, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.85, "rationale": "Autopilot addresses automated resource limit provisioning and wastage in cloud workloads.", "model": "claude-sonnet-5-5"}
++{"work_id": 2964142687, "categories": ["operating_systems", "computer_architecture"], "primary": "operating_systems", "confidence": 0.5, "rationale": "SSD data placement with deduplication is a storage-system problem that also touches flash hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400484472, "categories": ["hci", "software_engineering"], "primary": "hci", "confidence": 0.55, "rationale": "An augmented reality visitor guidance app, with a development methodology combining Mobile-D, RUP and BPM.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391306053, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Adaptive buffering of data transfers between a compute cloud and a storage cloud.", "model": "claude-sonnet-5-5"}
++{"work_id": 4292057429, "categories": ["privacy_preserving_computation", "distributed_computing"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "A robust private information retrieval scheme over distributed storage with network coding.", "model": "claude-sonnet-5-5"}
++{"work_id": 3027986714, "categories": ["operating_systems", "edge_computing"], "primary": "operating_systems", "confidence": 0.85, "rationale": "A new OS architecture that blends distributed OS, unikernel and LWK ideas for IoT offloading.", "model": "claude-sonnet-5-5"}
++{"work_id": 4254652521, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Optical disk storage hardware and servo systems, which is closest to digital storage hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204340147, "categories": ["cybersecurity", "operating_systems", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "Confidential VMs with AMD SEV for nested virtualization; hypervisor security against untrusted insiders, virtualization in clouds.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225872280, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Book chapter on Ansible modules; a practical tutorial on sysadmin tooling, not clearly research in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3189882660, "categories": [], "primary": null, "confidence": 0.9, "rationale": "VMware license price list; not research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4224674551, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Statistical distribution modelling with MLE and goodness-of-fit divergence; probabilistic modelling, applied to COVID data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406461284, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Characterization of read disturb in 3D NAND flash for storage reliability; memory hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4293193165, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Digital predistortion for power amplifiers is signal processing and communications hardware, which does not fit the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3033055137, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.93, "rationale": "File system design and implementation with performance evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3167972790, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.55, "rationale": "Uses NTFS features such as ADS, hard links and junctions to organize digital media; file system usage.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117899816, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Socio-political theoretical framework; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3095048882, "categories": ["operating_systems", "cloud_computing"], "primary": "operating_systems", "confidence": 0.7, "rationale": "I/O deduplication for primary storage, implemented in Linux, targeting cloud storage; storage systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4380875641, "categories": [], "primary": null, "confidence": 0.4, "rationale": "NILM appliance classification with feature selection and Bayesian optimization; smart grid signal-processing/ML application, not matching any ontology category clearly.", "model": "claude-sonnet-5-5"}
++{"work_id": 4213090496, "categories": ["operating_systems", "cloud_computing"], "primary": "operating_systems", "confidence": 0.85, "rationale": "Block IO control in the Linux kernel for containers in datacenters.", "model": "claude-sonnet-5-5"}
++{"work_id": 2902405086, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.9, "rationale": "In-kernel file system design (BetrFS) with efficient renames.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226300713, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.4, "rationale": "Practical chapter on Azure Virtual Desktop storage with FSLogix; cloud service usage.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401508056, "categories": ["operating_systems", "computer_architecture"], "primary": "operating_systems", "confidence": 0.7, "rationale": "ZNS SSD write handling with an indirection layer; storage systems and device firmware design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3173583453, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.7, "rationale": "Performance comparison of Windows 10 and Linux Mint operating systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 2901264075, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.7, "rationale": "SMR disk management scheme with address mapping and cleaning; storage systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 2973989512, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.45, "rationale": "Data placement algorithm for tiered storage using a Kohonen network; storage management.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386952147, "categories": ["distributed_computing", "algorithms"], "primary": "distributed_computing", "confidence": 0.7, "rationale": "Caching for low latency in distributed erasure-coded storage with sampling-based approximation and theoretical guarantees.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387067117, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Practical guide to persistence options in iOS app development.", "model": "claude-sonnet-5-5"}
++{"work_id": 3181207510, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Mobile app development for classroom use; education-oriented system, weak fit to any category, closest is HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 2894822167, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Benchmarking CernVM-FS, a distributed file system for software distribution, including concurrent publication.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401070015, "categories": ["ml_theory", "computational_biology"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Statistical method (independent vector analysis with hypothesis testing) applied to multi-task fMRI data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392711967, "categories": ["computer_architecture", "operating_systems"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Provisioning tool for computational storage drives with erasure coding offload; hardware and storage systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3097083941, "categories": ["software_engineering", "cloud_computing"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Design of an open-source, scalable online code execution system (Judge0).", "model": "claude-sonnet-5-5"}
++{"work_id": 3212917515, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.3, "rationale": "Education study using a custom RTOS; the main contribution is pedagogical, so the OS fit is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 7137550668, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.3, "rationale": "Lightning talk on efficient unpacking of software from CernVM-FS; very little information.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213061804, "categories": ["ai_accelerators", "llm"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "MoE offloading with prefetching, caching and graph runtime for efficient LLM inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408213108, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Book chapter fragment with a quote; no computer science research content discernible.", "model": "claude-sonnet-5-5"}
++{"work_id": 3170878721, "categories": ["operating_systems", "computer_architecture"], "primary": "operating_systems", "confidence": 0.7, "rationale": "Lightweight hypervisor with partitioning and scheduling for mixed-criticality IoT; virtualisation and real-time systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3008989608, "categories": ["database_systems", "distributed_computing"], "primary": "distributed_computing", "confidence": 0.55, "rationale": "Improves dataset upload to object storage via client-server cooperative merging and splitting of many small objects; a storage and data movement systems topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4224080142, "categories": ["operating_systems"], "primary": "operating_systems", "confidence": 0.9, "rationale": "Proposes a dynamic time quantum for round-robin CPU scheduling.", "model": "claude-sonnet-5-5"}
++{"work_id": 3193637602, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Comparative study of mobile piano-teaching apps (Flowkey and Yousician) for education; closest fit is the HCI category.", "model": "claude-sonnet-5-5"}
++{"work_id": 2981636369, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education-focused framework for learning styles and student performance; the abstract gives no clear computer-science contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394586023, "categories": ["cloud_computing", "distributed_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Trace analysis and optimization of container image management in a production Container Registry for cloud environments.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408867919, "categories": ["operating_systems", "database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Defragmentation and garbage collection for deduplicated backup storage; a storage systems topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 7137478203, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "CVMFS service architecture at CERN, covering stratum servers, S3 storage backend, and squid caches; distributed storage and middleware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200058389, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Classical and Bayesian estimation and confidence intervals for a process capability index; statistical methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3022441546, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Data science curriculum for school students and AI education; computing education, not covered by any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3113396090, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Nonlinear plasmonics physics in gold nanostructures; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3082840770, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.92, "rationale": "FPGA architecture optimization for area and reconfigurability.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112282291, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "In-storage hardware acceleration of hyperdimensional computing learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402895027, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "In-field testing of chip transition faults; hardware reliability and test (EDA).", "model": "claude-sonnet-5-5"}
++{"work_id": 3142919340, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.85, "rationale": "Clock distribution design for RSFQ circuits; VLSI/EDA.", "model": "claude-sonnet-5-5"}
++{"work_id": 3038664825, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Device physics of transistor degradation, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403361878, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.75, "rationale": "FPGA hardware design for sprite drawing with high-level synthesis, area and power tradeoffs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411019384, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Array signal processing DOA deception; not clearly within CS categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382052984, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.93, "rationale": "Survey of processor pipelines in RISC and CISC architectures.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414630583, "categories": ["ai_accelerators", "information_retrieval"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "CPU-NPU heterogeneous system for serving vector embeddings in RAG inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407949197, "categories": ["distributed_computing", "algorithms"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "System-level fault diagnosis algorithms for multiprocessor systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3040997825, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "Maps 3D CNN onto a vector DSP with Winograd optimisation; hardware-aware DNN acceleration.", "model": "claude-sonnet-5-5"}
++{"work_id": 4383502097, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.85, "rationale": "QCA digital circuit design of XOR gates and a code converter; VLSI/emerging logic design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400224459, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.92, "rationale": "Minimal-area RISC-V core on FPGA for IoT; microarchitecture.", "model": "claude-sonnet-5-5"}
++{"work_id": 2809334893, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Mixed-signal SoC with a compressive sensing encoder; hardware circuit design.", "model": "claude-sonnet-5-5"}
++{"work_id": 7206165071, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.85, "rationale": "Multi-gate transistor based adder/subtractor logic design; VLSI low-power design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2942318936, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.85, "rationale": "Model-based testing and quality assurance for component-based embedded applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 2966831047, "categories": ["distributed_computing", "ai_accelerators"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Hybrid data/model parallelism for multi-GPU DL training; distributed ML training systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3158474254, "categories": ["cybersecurity", "ai_safety"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "EM side-channel attack extracting BNN architecture and weights from FPGA hardware; security of AI systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160405187, "categories": ["formal_verification", "computational_complexity"], "primary": "formal_verification", "confidence": 0.5, "rationale": "ROBDD-based computation of automaton reactions and sequential circuit testing; automata and BDD methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133560921, "categories": ["computer_architecture", "cybersecurity"], "primary": "computer_architecture", "confidence": 0.88, "rationale": "CXL memory controller hardware for reliability and Row Hammer/side-channel protection.", "model": "claude-sonnet-5-5"}
++{"work_id": 2946432794, "categories": ["computer_architecture", "ml_theory"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Brain-inspired hyperdimensional clustering, mapped for processing-in-memory; the focus is on a memory-centric algorithm and architecture.", "model": "claude-sonnet-5-5"}
++{"work_id": 3206763996, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.92, "rationale": "Approximate compressors and multipliers: VLSI arithmetic circuit design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410057699, "categories": ["algorithms", "computer_architecture"], "primary": "algorithms", "confidence": 0.6, "rationale": "GPU-accelerated parallel graph partitioning algorithm for CAD/EDA.", "model": "claude-sonnet-5-5"}
++{"work_id": 3142700669, "categories": ["algorithms", "formal_verification"], "primary": "algorithms", "confidence": 0.4, "rationale": "Graph products for real-time process scheduling, specified in CSP; graph-theoretic and combinatorial analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399426830, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.35, "rationale": "Medallion data-engineering architecture applied to CPU benchmark analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386926746, "categories": ["cybersecurity", "computer_architecture"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Thermal covert channels in FPGA systems are a hardware security topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 3185457619, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.93, "rationale": "FPGA-based CNN hardware accelerator coprocessor design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2886646777, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Time synchronization between FlexRay and 802.1AS Ethernet networks in vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 3082922165, "categories": ["operating_systems", "algorithms"], "primary": "operating_systems", "confidence": 0.5, "rationale": "Real-time schedulability analysis and task models for scheduling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386764847, "categories": ["cybersecurity", "operating_systems"], "primary": "cybersecurity", "confidence": 0.9, "rationale": "Software covert channel attacks that exploit OS mutual exclusion and synchronization primitives.", "model": "claude-sonnet-5-5"}
++{"work_id": 2888777304, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Condensed-matter/spintronics materials measurement; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3147171055, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.55, "rationale": "Device-level transistor design and analog VLSI circuit benchmarking; closest to VLSI/hardware design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367663628, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.92, "rationale": "MRAM hybrid cache evaluation in RISC-V systems with prefetching strategy; microarchitecture and memory hierarchy.", "model": "claude-sonnet-5-5"}
++{"work_id": 2784201766, "categories": ["cybersecurity", "computer_architecture"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Memristor crossbar PUFs for hardware security; device-level hardware security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416176040, "categories": ["sustainable_computing", "llm", "ai_accelerators"], "primary": "sustainable_computing", "confidence": 0.88, "rationale": "Carbon footprint modeling of LLM training with hardware-aware, embodied and operational carbon accounting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415481175, "categories": ["computer_architecture", "sustainable_computing"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Integrated CMOS energy harvesting circuit for IoT nodes; VLSI circuit design with energy-harvesting angle.", "model": "claude-sonnet-5-5"}
++{"work_id": 3162223152, "categories": ["cybersecurity", "computer_architecture"], "primary": "cybersecurity", "confidence": 0.65, "rationale": "FPGA RO-PUF quality improvement using PCA; hardware security primitive on FPGA.", "model": "claude-sonnet-5-5"}
++{"work_id": 7170236789, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.92, "rationale": "Parallel iterative numerical solvers for PDEs (Helmholtz, fractional Poisson) with OpenMP.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416146873, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.45, "rationale": "Reproducibility report for HPC paper on GPU resilience; HPC systems context, minimal contribution itself.", "model": "claude-sonnet-5-5"}
++{"work_id": 3010845351, "categories": ["distributed_computing", "algorithms"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Mapping parallel HPDBSCAN clustering onto a modular exascale supercomputer architecture; HPC software.", "model": "claude-sonnet-5-5"}
++{"work_id": 2800869073, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Analog circuit synthesis and performance evaluation is electronic design automation, which falls under computer_architecture.", "model": "claude-sonnet-5-5"}
++{"work_id": 4213103670, "categories": ["cybersecurity"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "It exploits DVFS to build a covert channel, which is a hardware and system security attack.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392928691, "categories": ["ai_accelerators", "computer_vision"], "primary": "ai_accelerators", "confidence": 0.85, "rationale": "It designs an FPGA CNN accelerator for YOLOv3 object detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 2889574478, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "It profiles memory access patterns to design a tailored memory subsystem for application-specific hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 3023376085, "categories": ["programming_languages", "distributed_computing"], "primary": "programming_languages", "confidence": 0.6, "rationale": "It concerns programming BSP and Multi-BSP parallel algorithms in the ML language.", "model": "claude-sonnet-5-5"}
++{"work_id": 4256423899, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Fault detection and testing of analog circuits is a hardware design and test topic, so the fit to the ontology is weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 3199043212, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "It reduces test power for VLSI circuits using test pattern reordering, which is EDA and low-power design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3004719967, "categories": [], "primary": null, "confidence": 0.9, "rationale": "It is railway safety risk assessment, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2911751195, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.92, "rationale": "It presents an FPGA circuit design toolset for modelling heterogeneous FPGA architectures, which is reconfigurable computing and EDA.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312741045, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "It presents a compact model of phase change memory for circuit design and simulation, which is memory device and EDA modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411122680, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Railway signalling circuit simulation in Multisim; electrical engineering, not clearly computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416959540, "categories": ["scientific_computing", "algorithms"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Complexity analysis of certified numerical homotopy tracking with interval arithmetic; numerical methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415164777, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Control-theory data-driven predictive control; systems and control, not clearly in CS ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4235658738, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Numerical linear algebra: Thomas algorithm for tridiagonal systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387706174, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Convex optimization inertial dynamics convergence analysis; numerical optimization theory with numerical examples.", "model": "claude-sonnet-5-5"}
++{"work_id": 6940648869, "categories": [], "primary": null, "confidence": 0.7, "rationale": "LQR control design for power plant; control/power engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4224920451, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Data-driven symbolic regression for acoustic inverse problem and wave FEM speedup; scientific ML.", "model": "claude-sonnet-5-5"}
++{"work_id": 4363675873, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "Numerical B-spline Galerkin discretisation of a fractional PDE.", "model": "claude-sonnet-5-5"}
++{"work_id": 2969289037, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.55, "rationale": "Scientific data access software package for solar observatory data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4205760564, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Numerical computing chapter on linear algebra algorithms and implementations.", "model": "claude-sonnet-5-5"}
++{"work_id": 6967178263, "categories": ["computational_biology", "scientific_computing"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Evaluates machine-learned interatomic potentials for biomolecular simulations (proteins, water); molecular modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 3134502016, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Commentary on reproducibility and secondary analysis of datasets; not clearly CS research in a listed category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404306456, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.95, "rationale": "Virtual element discretization with error analysis for MHD equations.", "model": "claude-sonnet-5-5"}
++{"work_id": 7152139045, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.95, "rationale": "Domain decomposition and multiscale numerical methods, preconditioners for flood modelling PDEs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4327485924, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "Coupling PyTorch with Firedrake PDE framework; scientific machine learning software.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406816254, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Inventory and supply chain simulation-optimization; operations research, not within the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4376654240, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.93, "rationale": "Iterative numerical solver scheme for thermal radiative transfer PDEs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206279532, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applies a basic backpropagation network to railway passenger flow prediction; a routine application with no clear category fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 3199718208, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.92, "rationale": "Structure-preserving finite element discretization of stochastic thin-film equation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313597919, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.95, "rationale": "High-order DG spectral element method with shock capturing for compressible flow.", "model": "claude-sonnet-5-5"}
++{"work_id": 3210446440, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Bi-criteria optimization of public transport planning using local search and genetic algorithms with an ML oracle; combinatorial optimization/metaheuristics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405077071, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Numerical modelling and inverse-problem image reconstruction (Tikhonov, CG) for electromagnetic tomography parameter optimisation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155389303, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "Preconditioned conjugate gradient solvers for boundary element linear systems: numerical linear algebra.", "model": "claude-sonnet-5-5"}
++{"work_id": 4291941306, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Lie symmetry analysis of a PDE model is applied mathematics, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3216320654, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control theory for linear systems with parameter estimation; not clearly within a CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3084714138, "categories": ["computational_biology", "database_systems"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Workflow importing omics metadata from biological databases into data papers; bioinformatics data publishing and metadata standards.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408094557, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Tyre dynamics PDE modelling and stability analysis is mechanical engineering/applied math, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417347238, "categories": ["scientific_computing", "algorithms"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Optimization-based numerical algorithm for tensor rank decomposition; multilinear algebra computation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2963395393, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Pure mathematics on homogenization of elliptic PDEs; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382681213, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.7, "rationale": "Iterative numerical methods for polynomial zeros with convergence analysis; numerical analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163167038, "categories": ["scientific_computing", "distributed_computing"], "primary": "scientific_computing", "confidence": 0.9, "rationale": "Nonlinear domain decomposition preconditioners for Newton solvers of PDEs, with parallel implementation in Code_Aster.", "model": "claude-sonnet-5-5"}
++{"work_id": 3082621528, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.4, "rationale": "FEM-based numerical analysis of electrode topologies for bio-impedance; borderline computational engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 3012277325, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Numerical linear algebra: power method and convergence acceleration for dominant eigenvalues.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414593075, "categories": ["scientific_computing", "computer_vision", "deep_learning_architectures"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Physics-aware neural operator for 3D photoacoustic tomography inversion; scientific ML with imaging reconstruction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392110244, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "MHD simulation and equivalent heat source surrogate model with NN and genetic algorithm for arc thermal analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3157604584, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Mathematical analysis of swarming dynamics on a sphere; applied mathematics rather than computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394993168, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Epidemic-style mathematical model of corruption with optimal control; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2935049866, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Model order reduction techniques for discrete-time systems; numerical reduced-order models.", "model": "claude-sonnet-5-5"}
++{"work_id": 3159820714, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "LES computational modeling of flame flashback with tabulated chemistry; computational engineering simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3000672483, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Probabilistic modeling of constrained Ito processes via stochastic quantization, deriving Kalman filter variants.", "model": "claude-sonnet-5-5"}
++{"work_id": 7168156340, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.4, "rationale": "Adaptive filtering (RLS) for echo cancellation is signal processing, closest to audio enhancement, though the fit is loose.", "model": "claude-sonnet-5-5"}
++{"work_id": 3088835008, "categories": ["computer_networks"], "primary": "computer_networks", "confidence": 0.4, "rationale": "Beidou-GPS dual-mode positioning system for mobile communication devices; the fit with networking is loose.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390667245, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Semi-analytical numerical method for solving aggregation-breakage population balance equations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4322394327, "categories": ["computer_networks", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Co-simulation interfacing power system and ICT simulators; a simulation frameworks topic with a networking aspect.", "model": "claude-sonnet-5-5"}
++{"work_id": 2963298714, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.85, "rationale": "Sparse matrix storage and solution for finite element analysis, i.e. numerical linear algebra.", "model": "claude-sonnet-5-5"}
++{"work_id": 3123748159, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Collective intelligence study on stakeholder knowledge in fisheries; not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406141980, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Pure mathematics on existence of weak solutions to elliptic PDEs; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4310184906, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.8, "rationale": "RooFit performance optimisation through parallelization and batched computation in scientific software.", "model": "claude-sonnet-5-5"}
++{"work_id": 3116267422, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Inverse problem analysis for photoacoustic tomography with numerical simulations.", "model": "claude-sonnet-5-5"}
++{"work_id": 2810988304, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.55, "rationale": "Monte Carlo regression algorithm with convergence proof for optimal stopping of McKean-Vlasov diffusions; numerical methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414647119, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.95, "rationale": "Client sampling and allocation in multi-model federated learning under non-i.i.d. data.", "model": "claude-sonnet-5-5"}
++{"work_id": 2977976693, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Building physics and daylighting study of sky luminance; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4286224978, "categories": ["federated_learning", "ml_theory", "distributed_computing"], "primary": "ml_theory", "confidence": 0.6, "rationale": "One-shot distributed statistical learning with theoretical error-rate analysis; also proposes a federated variant, FedReBoot.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401880100, "categories": ["distributed_computing", "algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Decentralised asynchronous ADMM optimisation algorithm with equivalence proof, applied to phase retrieval.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403791687, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.95, "rationale": "Validation-free contribution assessment of participants in federated learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411551389, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.93, "rationale": "Adaptive local-epoch federated algorithm for non-IID data in the Internet of Vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385602851, "categories": [], "primary": null, "confidence": 0.93, "rationale": "Ionospheric and geophysical remote sensing analysis of a solar eclipse; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404238664, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.92, "rationale": "Weighted federated learning for heterogeneous data and models in edge computing, with convergence analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 2921727985, "categories": [], "primary": null, "confidence": 0.93, "rationale": "Radiometric estimation model for a Moon-based Earth observation instrument; remote sensing physics, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390986554, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.88, "rationale": "Reverse-auction incentive mechanism for federated learning with quality awareness and training-time prediction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404743912, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Federated learning applied to multi-site MRI image-to-image translation, comparing aggregation methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407950319, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Computing education study of challenge-based learning; not within the ontology categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415473342, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.75, "rationale": "Federated training for mask recognition, studying local epochs and client numbers.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416017477, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.95, "rationale": "Asynchronous, staleness-aware personalized FL with communication sparsification and convergence analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313889858, "categories": ["federated_learning", "computer_vision"], "primary": "federated_learning", "confidence": 0.8, "rationale": "FL framework for COVID-19 chest X-ray classification using DenseNet-169.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148322461, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Complex systems/physics study of predictability in cellular automata; not clearly a CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154578995, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Survey of federated learning applications in healthcare.", "model": "claude-sonnet-5-5"}
++{"work_id": 3215093560, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.85, "rationale": "PolSAR remote-sensing terrain classification using an improved convolutional autoencoder.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124262753, "categories": ["federated_learning", "ai_agents"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Holonic hierarchical distributed learning framework with model aggregation, similar to FL; agent-based framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408110066, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Survey of federated learning for Alzheimer's detection, in the biomedical health informatics domain.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213405981, "categories": ["federated_learning", "ml_theory"], "primary": "federated_learning", "confidence": 0.93, "rationale": "Proposes a federated optimization method for heterogeneous data with theoretical convergence analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413395786, "categories": ["federated_learning", "computational_biology"], "primary": "federated_learning", "confidence": 0.8, "rationale": "Compares federated, centralized and partially pre-trained classifiers for ECG atrial fibrillation detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405697278, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Vertical federated learning method for integrated IoT services.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131434230, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.55, "rationale": "Conceptual discussion of federated learning and big data for business decision making; weak contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400909753, "categories": ["federated_learning", "edge_computing", "database_systems"], "primary": "federated_learning", "confidence": 0.82, "rationale": "Federated trajectory recovery framework that is lightweight for edge devices and communication-efficient.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125126928, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.93, "rationale": "Mitigates catastrophic forgetting under temporal drift in FL via client-side experience replay.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401508498, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.93, "rationale": "Plugin for handling concept drift and personalization in federated learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415004105, "categories": ["federated_learning", "nlp"], "primary": "federated_learning", "confidence": 0.82, "rationale": "Empirical study of FedAvg vs FedProx with BERT for Chinese news text classification under non-IID data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414119830, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.75, "rationale": "FL-based digital twin framework with dynamic node selection, tested on a Raspberry Pi edge testbed.", "model": "claude-sonnet-5-5"}
++{"work_id": 3210955338, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.35, "rationale": "Loosely framed cyber-social computing metrics; proposes a cloud-edge federated learning architecture only in part.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163589765, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Educational study of AI-supported instruction; weakly human-AI interaction, mostly education research rather than CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403284007, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Title indicates communication-efficient federated learning; abstract is missing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3007895144, "categories": ["ml_theory", "algorithms"], "primary": "ml_theory", "confidence": 0.75, "rationale": "Robust distribution estimation from untrusted batches with optimal polynomial-time estimator; statistical learning theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 7084124479, "categories": ["federated_learning", "ml_theory"], "primary": "federated_learning", "confidence": 0.85, "rationale": "New federated optimization method with convergence proof against FedAvg and FedProx.", "model": "claude-sonnet-5-5"}
++{"work_id": 3210763747, "categories": ["federated_learning", "ai_safety", "cybersecurity"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Client-side defense against model poisoning in FL with certified robustness guarantees.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204715907, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Qualitative education study of student perceptions of AI; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2907334421, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Power electronics wireless power transfer for a lunar rover; electrical engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401891542, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Federated adversarial method with adaptive aggregation for fault diagnosis under heterogeneous conditions.", "model": "claude-sonnet-5-5"}
++{"work_id": 4376149513, "categories": ["federated_learning", "edge_computing"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Survey of heterogeneity-aware federated learning for resource-constrained devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 2908566518, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Essay on an AI cosmology as a conceptual framing for the field; no clear category fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 7106710552, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Personalized federated test-time adaptation with balanced batch normalization for class-imbalanced clients.", "model": "claude-sonnet-5-5"}
++{"work_id": 7143299875, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Community service training on using Google Drive for teachers; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390204018, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Incentive and collaboration mechanisms for cross-silo federated learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4254855164, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Generic book chapter introduction with no clear research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313919668, "categories": ["federated_learning", "ml_theory"], "primary": "federated_learning", "confidence": 0.7, "rationale": "Federated learning of multiple Bayesian networks for multi-site data, using ADMM and structure learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412004712, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.75, "rationale": "Review of federated learning with no-code AI for BFSI data security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393928184, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.92, "rationale": "Client selection in FL using data entropy and model divergence to cut communication.", "model": "claude-sonnet-5-5"}
++{"work_id": 7085089513, "categories": ["federated_learning", "deep_learning_architectures", "edge_computing"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Federated graph learning with edge-client collaboration and missing-topology handling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381744345, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.88, "rationale": "Robust, communication-compressed federated learning framework against Byzantine clients.", "model": "claude-sonnet-5-5"}
++{"work_id": 4320855546, "categories": ["federated_learning"], "primary": "federated_learning", "confidence": 0.9, "rationale": "Diffusion strategy for non-IID FL with communication efficiency in 6G.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403353453, "categories": ["computer_vision", "ml_theory"], "primary": "computer_vision", "confidence": 0.4, "rationale": "Zero-shot learning with image classification and visual-aware knowledge graph completion; the focus is on knowledge representation and zero-shot learning. Vision is the closest fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128377297, "categories": ["foundation_models", "deep_learning_architectures"], "primary": "foundation_models", "confidence": 0.85, "rationale": "Time series foundation model study with scaling, data composition and ablations of a standard patch Transformer.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402260667, "categories": ["computer_vision", "foundation_models"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Remote sensing canopy height estimation comparing a geospatial foundation model with a U-Net.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169519287, "categories": ["foundation_models", "generative_ai", "reinforcement_learning", "ai_agents"], "primary": "foundation_models", "confidence": 0.3, "rationale": "Introductory overview chapter spanning ML paradigms, generative models, transformers and agents; broad and not specific.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402145393, "categories": ["foundation_models"], "primary": null, "confidence": 0.25, "rationale": "Perspective on practical AI in Earth sciences; mostly domain-level and only touches on foundation models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410550951, "categories": ["foundation_models", "deep_learning_architectures"], "primary": "foundation_models", "confidence": 0.6, "rationale": "Survey of self-supervised representation learning for EEG, including pre-trained foundation models for EEG.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171486454, "categories": ["foundation_models", "computational_biology"], "primary": "foundation_models", "confidence": 0.5, "rationale": "Self-supervised masked autoencoder pre-training on large unlabelled EEG/HRV data, transferred to HIE classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 7116294541, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Compares continual and transfer learning for concept drift in building thermal models and proposes a continual learning strategy; applied energy domain.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416183290, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Satellite image segmentation of glacial lakes using SAM 2.1.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133207779, "categories": ["foundation_models", "edge_computing"], "primary": "foundation_models", "confidence": 0.8, "rationale": "Pre-trained health foundation model on wearable data with continual parameter-efficient fine-tuning; targets memory-efficient edge deployment.", "model": "claude-sonnet-5-5"}
++{"work_id": 3189938493, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Transfer learning for small-sample signal modulation recognition; the venue suggests a communications application, but no category fits well, so deep learning is the loose fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412605435, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Online multi-source transfer learning for fault diagnosis; a task-specific ML method with no clearly matching application category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7172007151, "categories": ["edge_computing", "computational_biology"], "primary": "edge_computing", "confidence": 0.5, "rationale": "Meta-learning for efficient training and updating on wearable devices with reduced energy use, applied to seizure and AF detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404035916, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Cross-domain aspect sentiment triplet extraction using prompts; a classical NLP information extraction task.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171698092, "categories": ["foundation_models", "ai_safety"], "primary": "foundation_models", "confidence": 0.6, "rationale": "Robustness metric for pathology foundation models that evaluates shortcut learning and confounder sensitivity in their representations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417300144, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.55, "rationale": "Trunk-branch adaptive architecture for conflicting time series tasks, with expressivity and convergence theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 7130402983, "categories": ["llm", "ai_agents", "reinforcement_learning"], "primary": "llm", "confidence": 0.9, "rationale": "GLM-5 foundation language model with agentic and coding capabilities and asynchronous RL post-training.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389821395, "categories": ["reinforcement_learning", "multimodal_ai"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Uses CLIP vision-language models as reward sources to train RL agents on language goals.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203492078, "categories": ["foundation_models", "multimodal_ai"], "primary": "foundation_models", "confidence": 0.7, "rationale": "EEG foundation model with masked pretraining and instruction tuning, using a Q-Former and text embeddings for cross-dataset decoding.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386106645, "categories": ["foundation_models", "llm", "ai_accelerators"], "primary": "foundation_models", "confidence": 0.7, "rationale": "Open-source toolkit suite for efficient training, tuning, compression and inference of big pre-trained models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407186881, "categories": ["foundation_models"], "primary": "foundation_models", "confidence": 0.7, "rationale": "Adapts the pre-trained TabPFN in-context model via encoder fine-tuning and bagging; a foundation-model adaptation method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387929345, "categories": ["nlp"], "primary": "nlp", "confidence": 0.85, "rationale": "Few-shot semi-supervised classification of disaster tweets; a text classification task.", "model": "claude-sonnet-5-5"}
++{"work_id": 4315588764, "categories": ["foundation_models", "computational_biology"], "primary": "foundation_models", "confidence": 0.75, "rationale": "Self-supervised pretrained foundation model for EHR time-to-event prediction, with transfer evaluation and a model release.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167906451, "categories": ["llm", "foundation_models"], "primary": "llm", "confidence": 0.85, "rationale": "Continual parameter-efficient LoRA fine-tuning of 7-8B LLMs.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140202126, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Benchmark of PPG-based clinical prediction using deep learning; biomedical health informatics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412420015, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Few-shot partial multi-label learning methodology addressing label noise; general ML methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385757412, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.6, "rationale": "Semi-supervised personalized OSA detection from ECG; biomedical informatics method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390871840, "categories": ["ai_safety", "computer_vision"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Studies inheritance and correction of bias from pretrained models via finetuning data; a fairness focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 3160088594, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.85, "rationale": "Few-shot COVID-19 CT diagnosis using domain adaptation with synthetic images; medical image analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165377962, "categories": ["foundation_models"], "primary": "foundation_models", "confidence": 0.7, "rationale": "Evaluates time series foundation models, zero-shot versus fine-tuned with covariates, for building energy forecasting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399458826, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.9, "rationale": "Silhouette-based gait recognition with domain adaptation and a new dataset; unimodal vision.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165063713, "categories": ["foundation_models", "deep_learning_architectures"], "primary": "foundation_models", "confidence": 0.45, "rationale": "Multi-source transfer learning via sequential parameter merging and fine-tuning; transfer/continual learning focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313024900, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Unsupervised domain adaptation for radar activity classification using Margin Disparity Discrepancy, a theory-inspired method. Radar sensing is not clearly covered by any other category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214033248, "categories": ["foundation_models", "deep_learning_architectures"], "primary": "foundation_models", "confidence": 0.6, "rationale": "Adapts a frozen pretrained vision model (MAE) to time-series forecasting; transfer from a large pretrained model with lightweight adaptation modules.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138441586, "categories": ["foundation_models", "llm"], "primary": "foundation_models", "confidence": 0.85, "rationale": "Parameter-efficient fine-tuning (LoRA/DoRA) methods for adapting large pretrained models, evaluated on NLU and generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402355991, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "VAE and clustering representation learning for metro noise signals; applied self-supervised representation learning on audio-like signals.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391766533, "categories": ["computer_vision", "foundation_models"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Self-supervised pretraining of transformers on whole slide images for computational pathology; medical image analysis and visual representation learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405173494, "categories": ["ai_agents", "reinforcement_learning"], "primary": "ai_agents", "confidence": 0.75, "rationale": "Retrieval-augmented generalist agent acting in-context in new environments, with a transformer policy; sits between agent architectures and policy learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7130707312, "categories": ["foundation_models", "deep_learning_architectures"], "primary": "foundation_models", "confidence": 0.85, "rationale": "EEG foundation model with a residual vector quantization tokenizer and autoregressive pretraining on a large corpus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403992746, "categories": ["foundation_models"], "primary": "foundation_models", "confidence": 0.4, "rationale": "Selecting pre-trained models for transfer learning using meta-features; the abstract is minimal, so the judgment rests on the title.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413751397, "categories": ["computational_biology", "speech_audio", "multimodal_ai"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Survey of deep learning on speech, text and handwriting for cognitive decline detection; a biomedical health informatics application, with some multimodal and speech content.", "model": "claude-sonnet-5-5"}
++{"work_id": 3134294021, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "English event detection across data sources using transfer learning and adversarial training; a text mining task.", "model": "claude-sonnet-5-5"}
++{"work_id": 3139784666, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.3, "rationale": "Transfer of ML models between smart environments, with a knowledge-graph framework for describing models; loosely fits IoT and smart-building services.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167230086, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "CLIP-based multilingual zero-shot text classification that uses visual prototypes and a new benchmark.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417127389, "categories": ["multimodal_ai", "deep_learning_architectures", "foundation_models"], "primary": "multimodal_ai", "confidence": 0.65, "rationale": "Pruning combined with LoRA for efficient fine-tuning of CLIP vision-language models, evaluated on image-text retrieval.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167035799, "categories": ["foundation_models", "computer_vision"], "primary": "foundation_models", "confidence": 0.7, "rationale": "Endoscopy foundation model pretrained on 5.6M images and evaluated for gastric cancer analysis on medical images.", "model": "claude-sonnet-5-5"}
++{"work_id": 6966932760, "categories": ["generative_ai", "foundation_models"], "primary": "generative_ai", "confidence": 0.55, "rationale": "Dynamic synthetic data generation framework that improves scaling laws, evaluated on ImageNet.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161153937, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Studies how pre-processing and imaging equipment affect CNN performance for breast cancer detection in mammograms.", "model": "claude-sonnet-5-5"}
++{"work_id": 3126536330, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.4, "rationale": "Domain adaptation network for bearing fault diagnosis; an engineering application of transfer learning with no matching category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140216482, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.75, "rationale": "Factored norm and fused Triton GPU kernels that make DoRA fine-tuning faster and lighter in memory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312223859, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Side-channel countermeasure design using channel capacity; cryptographic implementation security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391216411, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Power system capacity allocation; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391582496, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Chaos/DNA-based image encryption scheme with S-box construction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4280508268, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.95, "rationale": "Cryptanalysis of the TinyJAMBU lightweight cipher: distinguishers and key recovery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4360859090, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Image encryption algorithm built as an SP-network with chaotic keys and DNA operations.", "model": "claude-sonnet-5-5"}
++{"work_id": 3116994096, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.95, "rationale": "Public-key encryption and signature scheme on twisted Edwards curves, with provable security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400493258, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.92, "rationale": "S-box construction for symmetric ciphers.", "model": "claude-sonnet-5-5"}
++{"work_id": 7199797550, "categories": ["applied_cryptography", "privacy_preserving_computation"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Thesis on cryptographic primitives for e-voting, with a homomorphic encryption prototype.", "model": "claude-sonnet-5-5"}
++{"work_id": 7016138925, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Side-channel security evaluation, masking and shuffling countermeasures for cryptographic implementations.", "model": "claude-sonnet-5-5"}
++{"work_id": 3176687530, "categories": ["privacy_preserving_computation", "blockchain", "cloud_computing"], "primary": "privacy_preserving_computation", "confidence": 0.7, "rationale": "Blockchain-assisted privacy-preserving EHR sharing in the cloud, with a privacy-utility objective.", "model": "claude-sonnet-5-5"}
++{"work_id": 3135040192, "categories": ["applied_cryptography", "ai_accelerators"], "primary": "applied_cryptography", "confidence": 0.55, "rationale": "GPU-accelerated entropy estimation (NIST SP 800-90B IID tests) for RNG evaluation; crypto-related with GPU parallel implementation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4255230922, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Side-channel CPA attack improved with variational mode decomposition; cryptanalysis/side-channel.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312999490, "categories": ["cloud_computing", "cybersecurity", "quantum_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Survey of cloud security and privacy concerns and quantum computing threat.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127423155, "categories": ["blockchain", "applied_cryptography"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain-based record authentication framework with DIDs, VCs, IPFS and Merkle audit trails.", "model": "claude-sonnet-5-5"}
++{"work_id": 4213389055, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.65, "rationale": "Side-channel attack improvement against AES S-boxes with genetic algorithm comparison; cryptanalysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 7108084985, "categories": ["applied_cryptography", "quantum_computing"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Chapter on quantum random number generators for cryptographic randomness.", "model": "claude-sonnet-5-5"}
++{"work_id": 4246586608, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Symmetric-key encryption algorithm using genetic crossover and mutation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401369144, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Operations research/economics model of product repositioning; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413953789, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Practical implementation of isogeny-based post-quantum cryptography; abstract is publisher boilerplate but title is clear.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401634049, "categories": ["applied_cryptography", "blockchain", "privacy_preserving_computation"], "primary": "applied_cryptography", "confidence": 0.75, "rationale": "Threshold blind signature scheme on Paillier cryptosystem applied to blockchain, giving anonymity and key management.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011920068, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Constraint codes plus ECC for NAND flash reliability is a memory-systems and coding problem; closest fit is computer_architecture.", "model": "claude-sonnet-5-5"}
++{"work_id": 2963804406, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.95, "rationale": "Improves the security of the lattice-based GGH public-key scheme.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200321003, "categories": ["applied_cryptography", "cybersecurity"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Correlation power analysis side-channel attack on the SPECK cipher, a cryptanalysis and implementation-security topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401694671, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.55, "rationale": "Cloud security for healthcare images using an optimized LSTM classifier on image pixels.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215253795, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.92, "rationale": "Blockchain-based credential issuance and verification using smart contracts, DIDs and IPFS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2922278162, "categories": ["applied_cryptography", "software_engineering"], "primary": "applied_cryptography", "confidence": 0.5, "rationale": "Chat application that encrypts stored messages with Caesar cipher and 3DES; the focus is on applying cryptography.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417415103, "categories": ["federated_learning", "blockchain", "privacy_preserving_computation"], "primary": "federated_learning", "confidence": 0.88, "rationale": "Combines federated learning, blockchain, differential privacy and homomorphic encryption for medical data sharing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3157885010, "categories": ["applied_cryptography", "computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "Hardware modular reduction and memory access design for the NTT used in post-quantum cryptography; area-efficient VLSI work.", "model": "claude-sonnet-5-5"}
++{"work_id": 4251764918, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Overview chapter on hash functions, public-key cryptography and PGP.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416234071, "categories": ["applied_cryptography", "computational_complexity"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Bounds on monomial density of Boolean functions, used for a probabilistic degree test and applied to stream ciphers.", "model": "claude-sonnet-5-5"}
++{"work_id": 2785483644, "categories": ["cybersecurity", "privacy_preserving_computation"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Biometric template protection via fuzzy vault; chaff generation speedup, security of biometric systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3126856290, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Power systems Volt/VAR control; electrical engineering, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3204643661, "categories": ["privacy_preserving_computation", "applied_cryptography"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Single-server private information retrieval using homomorphic encryption.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226498265, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Agricultural irrigation control with MPC; control engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405179471, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Historical biography of cryptographers; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164908005, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.8, "rationale": "Secret sharing scheme for key management; cryptographic construction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404987385, "categories": ["quantum_computing", "applied_cryptography"], "primary": "quantum_computing", "confidence": 0.7, "rationale": "Source-independent quantum random number generator with SDP security proof; quantum information processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391590938, "categories": ["privacy_preserving_computation", "ai_accelerators"], "primary": "privacy_preserving_computation", "confidence": 0.75, "rationale": "GPU-accelerated MPC framework for private ML training and Transformer inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392426149, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Secure two-party computation protocols for activation functions in privacy-preserving neural networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391096928, "categories": ["blockchain", "applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Survey of post-quantum digital signatures with advanced features for blockchains.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401247688, "categories": ["applied_cryptography", "cloud_computing"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Proof of retrievability scheme with provable security for integrity auditing of outsourced storage; cryptographic construction, cloud storage context.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200232553, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Discusses NIST post-quantum cryptography standardization and transition.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160766211, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "Unifying framework for MPC protocols with dynamic committees; MPC as a cryptographic primitive.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220773489, "categories": ["applied_cryptography", "edge_computing"], "primary": "applied_cryptography", "confidence": 0.6, "rationale": "Performance evaluation of SPECK block cipher on Raspberry Pi, a lightweight cryptography implementation study.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402836518, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.7, "rationale": "Chaos and DNA coding based image encryption scheme; cryptographic design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4378805032, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Power systems frequency control with EV stations and MPC (model predictive control); not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3073667118, "categories": ["applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.9, "rationale": "Key exchange protocol over a non-commuting group with security proof based on an NP-complete problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 3090466325, "categories": ["applied_cryptography", "algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Algorithmic construction of finite field elements of provably high order, motivated by cryptography and coding.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402426511, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.9, "rationale": "Thesis on quantum key distribution and bosonic quantum error-correcting codes.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392502209, "categories": ["applied_cryptography", "ai_accelerators"], "primary": "applied_cryptography", "confidence": 0.75, "rationale": "GPU-parallel optimized implementation of the SPHINCS+ post-quantum signature scheme.", "model": "claude-sonnet-5-5"}
++{"work_id": 3115113005, "categories": ["computational_complexity", "computer_graphics", "compilers"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Introduces a computational model abstracting fragment-shader machines, with running time analysis and transcompilation to shader programs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4232616731, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.35, "rationale": "Data visualization and information design overview; visualisation is listed under graphics, though the item is essentially an editorial.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381744354, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "CSCW design method and data visualization dashboard for a business intelligence platform.", "model": "claude-sonnet-5-5"}
++{"work_id": 3185991883, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Image-based rendering and view synthesis from light fields for non-Lambertian objects.", "model": "claude-sonnet-5-5"}
++{"work_id": 2909271333, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.45, "rationale": "Formalization of geometric object configurations for packing, layout and covering problems, a combinatorial optimisation topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413330002, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.75, "rationale": "Workshop panel on rendering needs for esports, concerning rendering research and latency.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382541407, "categories": ["distributed_computing", "scientific_computing"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "MPI+X hybrid parallelization and dynamic load balancing of an unstructured DSMC solver on HPC systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285816232, "categories": ["computer_graphics", "distributed_computing"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Parallel-processing rendering acceleration for complex 3D scenes.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400878583, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.7, "rationale": "3D modelling from hand-drawn strokes with spline fitting and color rendering for graphic design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035417602, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Design of a public service advertisement video; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7139845931, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Algorithms for reconstructing Euclidean distance matrices, with a cipher key-transfer application; mainly mathematical algorithm design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3044538714, "categories": ["computer_graphics", "computer_vision"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Neural scene representation for fast free-viewpoint rendering (NeRF variant).", "model": "claude-sonnet-5-5"}
++{"work_id": 4249339162, "categories": ["scientific_computing", "computer_graphics"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Mesh improvement methods (Laplacian smoothing, optimization) for mesh quality, in a numerical meshing context.", "model": "claude-sonnet-5-5"}
++{"work_id": 2758763654, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Augmented reality drum visualization app with marker tracking tested under camera resolution and angle parameters.", "model": "claude-sonnet-5-5"}
++{"work_id": 2785914471, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Information visualization: assessing clutter reduction in parallel coordinates using image processing measures.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414732984, "categories": ["generative_ai", "multimodal_ai", "speech_audio"], "primary": "generative_ai", "confidence": 0.6, "rationale": "LSTM model generating chords from image colors via emotion quantification; audio-visual generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389429756, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Artwork using AI-generated portraits grounded in Mangaian cosmology; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7151078784, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.7, "rationale": "Performance analysis of quantized easing functions for game animation in a game engine.", "model": "claude-sonnet-5-5"}
++{"work_id": 4229820775, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Historical narrative book on Xbox and DirectX; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417339164, "categories": ["computer_vision", "computer_graphics"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Neural distance field for 3D reconstruction and novel-view synthesis of dynamic urban scenes.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409753672, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Visual identity/graphic design project for a local art market; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7163058269, "categories": ["hci", "computer_graphics"], "primary": "hci", "confidence": 0.6, "rationale": "Interactive visual analytics strategies coordinating process mining views; human-centered interaction and visualization.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382563412, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "Prototype in Unity game engine recreating non-human sensory perspective; game engine and audio-visual representation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2930742311, "categories": ["scientific_computing", "computer_graphics"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "SPH multiphase fluid simulation of landslide waves with visualization; numerical simulation of a physical hazard.", "model": "claude-sonnet-5-5"}
++{"work_id": 2340447824, "categories": ["computer_graphics", "computer_architecture"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Parallel hardware architecture for video-rate volume rendering via ray casting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404698773, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Examines AI text-to-video platforms (Runway, Luma) and analyzes their generated moving images.", "model": "claude-sonnet-5-5"}
++{"work_id": 3183006928, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.85, "rationale": "Rendering and scientific visualization application built on OSPRay ray tracing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167851020, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Production session on film animation and lighting pipeline at Pixar.", "model": "claude-sonnet-5-5"}
++{"work_id": 7206470547, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Title and abstract give no usable information (a page count only); cannot assign a category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400958279, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.95, "rationale": "Real-time stereo rendering reflection technique.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404548282, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.9, "rationale": "Real-time GPU rendering of raindrops on a virtual camera lens.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412588572, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Procedural pattern editing with interactive optimization, SIGGRAPH paper, validated with a user study.", "model": "claude-sonnet-5-5"}
++{"work_id": 2966003478, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.85, "rationale": "Evaluates mesh polygon reduction effects on 3D geometry, which is geometry processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7218025575, "categories": ["algorithms", "computer_graphics"], "primary": "algorithms", "confidence": 0.6, "rationale": "Compares maze generation algorithms for procedural content generation in games, with a solver-based complexity measure.", "model": "claude-sonnet-5-5"}
++{"work_id": 4375835151, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "3D promotional animation video produced with modeling and rendering software; weak research contribution, but it is graphics content creation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4280553375, "categories": ["computer_vision", "computer_graphics"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Fisheye image stitching for panorama reconstruction with mesh alignment and CUDA acceleration.", "model": "claude-sonnet-5-5"}
++{"work_id": 3037385141, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Spatial risk data visualization approach.", "model": "claude-sonnet-5-5"}
++{"work_id": 2789701916, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.75, "rationale": "Streamgraph visualization technique for exploring hierarchical time series, published in TVCG.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392521761, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.5, "rationale": "3D animation of a ventilator built in Autodesk Maya using the MDLC method; it is graphics content production.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221345862, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.8, "rationale": "Multiprojector image distortion correction for curved VR screens using reverse ray tracing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4296604574, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.4, "rationale": "Reflections on teaching visualization; visualization education spans graphics and HCI, with weak fit to either.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393211575, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.6, "rationale": "Topological data structures for geometric modeling (simplicial and cellular structures).", "model": "claude-sonnet-5-5"}
++{"work_id": 4385490078, "categories": ["computer_graphics"], "primary": "computer_graphics", "confidence": 0.95, "rationale": "Non-photorealistic rendering with suggestive contours for conveying shape.", "model": "claude-sonnet-5-5"}
++{"work_id": 4321788308, "categories": ["computer_graphics", "database_systems"], "primary": "computer_graphics", "confidence": 0.45, "rationale": "Interactive visualization system for supermarket sales data and item associations; touches on data mining.", "model": "claude-sonnet-5-5"}
++{"work_id": 2911494573, "categories": ["ai_accelerators", "scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "GPU SpMV auto-tuner is a numerical kernel optimization; it is more about sparse linear algebra than ML hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405709446, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.6, "rationale": "AIGC for digital humans in the metaverse, covering content generation and multimodal interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4220934601, "categories": ["computer_graphics", "hci"], "primary": "computer_graphics", "confidence": 0.45, "rationale": "VR and 3D modeling for print advertisement visual optimization, using image mining.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417539761, "categories": ["computer_vision", "computer_graphics"], "primary": "computer_vision", "confidence": 0.8, "rationale": "3D reconstruction of articulated objects with Gaussian splatting and a part-aware representation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2804059852, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Phase error correction for fringe projection 3D profilometry, a 3D reconstruction problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417124644, "categories": ["generative_ai", "computer_graphics", "computer_vision"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Generative 4D inbetweening using video interpolation, diffusion and 3DGS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131893792, "categories": ["multimodal_ai", "computer_vision", "robotics"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Uses LLaVA VLM with prompt engineering for pedestrian behavior detection from a robot's view.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160568041, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Benchmark for chart understanding by VLMs, including hallucination detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417286526, "categories": ["multimodal_ai", "computer_vision"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Explores VLMs with prompting for driver monitoring.", "model": "claude-sonnet-5-5"}
++{"work_id": 3157008771, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.75, "rationale": "Statistical methods for ranking data: rank correlation, decision trees, bagging and boosting.", "model": "claude-sonnet-5-5"}
++{"work_id": 3212108239, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Continual learning of language grounding using CLIP embeddings for referring expressions.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204869979, "categories": ["multimodal_ai", "reinforcement_learning"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Active manga VQA with group-relative policy optimization of a multimodal model.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402703549, "categories": ["multimodal_ai", "foundation_models"], "primary": "multimodal_ai", "confidence": 0.9, "rationale": "Open family of large multimodal models with datasets and training recipe.", "model": "claude-sonnet-5-5"}
++{"work_id": 7083693837, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Benchmark evaluating multimodal clinical reasoning of VLMs in neurology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4324302630, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Applies CLIP-style multimodal models to digital humanities collections.", "model": "claude-sonnet-5-5"}
++{"work_id": 7172396779, "categories": ["multimodal_ai", "computer_vision"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Training-free text-to-video state moment retrieval with a frozen image-text encoder.", "model": "claude-sonnet-5-5"}
++{"work_id": 3176349206, "categories": ["multimodal_ai", "deep_learning_architectures"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Cross-modal retrieval using adversarial networks and GNNs for image-text common representations.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220827782, "categories": ["information_retrieval", "multimodal_ai"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Multimodal retrieval in a RAG system for pediatric nutrition care; the abstract is uninformative, so the label rests on the title.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412945255, "categories": ["multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Studies compositional generalization of multimodal LLMs on medical imaging and releases the Med-MAT dataset.", "model": "claude-sonnet-5-5"}
++{"work_id": 7197818024, "categories": ["computer_vision", "computational_biology"], "primary": "computer_vision", "confidence": 0.55, "rationale": "Thesis on multimodal deep-learning analysis of medical data for diabetic retinopathy classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415538349, "categories": ["multimodal_ai", "nlp", "llm"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "MLLM with dialogue graphs for multimodal conversational aspect-based sentiment analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403884255, "categories": ["multimodal_ai", "llm", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Open-source multimodal model trained as an evaluator (LMM-as-a-judge) that also gives reward signals for preference learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134939474, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Multimodal sentiment analysis with cross-modal attention fusion over text, audio and visual inputs.", "model": "claude-sonnet-5-5"}
++{"work_id": 3205782068, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Audio-visual scene-aware dialog task, dataset and audio-visual transformer system with temporal reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166830964, "categories": ["multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Evaluates large vision-language models on Bangla medical visual question answering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394564345, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Visual dialog model with dynamic cross-modal interaction and graph reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381956051, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Multimodal text-image sentiment analysis using a vision-language pretrained model and contrastive learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214149876, "categories": ["multimodal_ai", "hci"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Benchmarks MLLMs for chart-to-data/design extraction and presents a mixed-initiative interactive system.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160273623, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Multimodal corpus of audio, video and motion with language annotations, plus speech-action alignment algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416060734, "categories": ["multimodal_ai", "generative_ai"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Diffusion model that completes missing modality features for multimodal emotion and intent recognition.", "model": "claude-sonnet-5-5"}
++{"work_id": 7106312846, "categories": ["multimodal_ai", "reinforcement_learning"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Chest X-ray report generation from images with CLIP encoders and a REINFORCE reward.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414169798, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Medical image captioning and report generation from chest X-rays with CheXNet and BioGPT.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117449861, "categories": ["multimodal_ai", "foundation_models"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Multi-hop visual chain-of-thought VQA on a frozen vision-language model with low-rank adapters.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387968461, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.88, "rationale": "Knowledge-enhanced image captioning using a vision-language pre-training model, with a new benchmark.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131242786, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.9, "rationale": "Comparative evaluation of text-to-image diffusion models with image quality metrics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155044497, "categories": ["llm", "multimodal_ai", "speech_audio", "reinforcement_learning"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "RL training of multimodal LLMs for speech summarization directly from audio.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410698082, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Multimodal sentiment analysis fusing text, audio and visual data via cross-modal attention and contrastive learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128505395, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Video-text alignment and reasoning to generate explanations for fake news videos.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387969592, "categories": ["generative_ai", "multimodal_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Stable Diffusion-based image reconstruction from fMRI using CLIP text and visual features.", "model": "claude-sonnet-5-5"}
++{"work_id": 2998378979, "categories": ["multimodal_ai", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Cross-modal retrieval across multiple modalities with a graph attention network and a GAN.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404688739, "categories": ["multimodal_ai", "ai_safety"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Interpretability of image-text pairing in multimodal data fusion using a counterfactual method.", "model": "claude-sonnet-5-5"}
++{"work_id": 3201123210, "categories": ["multimodal_ai", "deep_learning_architectures"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Multimodal emotion recognition from text, speech and vision using a hierarchical fusion graph convolutional network.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412082472, "categories": ["multimodal_ai", "deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.65, "rationale": "Token pruning for efficient inference of large multimodal models such as BLIP-2.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412438803, "categories": ["multimodal_ai", "generative_ai"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Poster layout generation using multimodal training and chain-of-thought.", "model": "claude-sonnet-5-5"}
++{"work_id": 3034253961, "categories": ["ai_agents", "multimodal_ai", "reinforcement_learning"], "primary": "ai_agents", "confidence": 0.75, "rationale": "Vision-and-language navigation agent trained with imitation learning and curriculum RL, with new benchmarks.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213885858, "categories": ["multimodal_ai"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Robust image-text crisis classification that handles degraded or missing modalities.", "model": "claude-sonnet-5-5"}
++{"work_id": 7126829820, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.8, "rationale": "Multimodal meme sexism classification with image and text embeddings and contrastive learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155094944, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Abstract visual reasoning on RAVEN with a dual-path attention model and contrastive rule learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169863823, "categories": ["multimodal_ai", "ai_agents"], "primary": "multimodal_ai", "confidence": 0.7, "rationale": "Multimodal LLM adapted for vision-language navigation, with a Q-Former, LoRA and Ascend NPU training.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313013123, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.65, "rationale": "Survey of multimodal event detection from Twitter data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404450837, "categories": ["multimodal_ai", "generative_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.85, "rationale": "Any-to-many multimodal LLM that generates text plus multiple other modalities.", "model": "claude-sonnet-5-5"}
++{"work_id": 4283511515, "categories": ["multimodal_ai", "nlp"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Multimodal cheapfake detection using image captioning to relate images and captions.", "model": "claude-sonnet-5-5"}
++{"work_id": 4398157060, "categories": ["ai_agents", "multimodal_ai", "llm"], "primary": "ai_agents", "confidence": 0.75, "rationale": "LLM-based vision-and-language navigation agent with a memory map and chain-of-thought reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405779000, "categories": ["multimodal_ai", "computer_vision"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Uses the language modality to guide vision anomaly detection through cross-modal guidance.", "model": "claude-sonnet-5-5"}
++{"work_id": 7114776455, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Template-based text-to-image generation of accessible illustrations for simplified text.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171869391, "categories": ["multimodal_ai", "llm", "information_retrieval"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Retrieval-based knowledge augmentation of multimodal LLMs for substation inspection, with a benchmark.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118751186, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Library science user survey of digital library services; not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4321485607, "categories": ["software_engineering", "information_retrieval"], "primary": "software_engineering", "confidence": 0.8, "rationale": "Code search framework combining IR, bi-encoder and cross-encoder ranking to aid developers.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124998404, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Recommender system for streams and colleges with ML and an NLP interface.", "model": "claude-sonnet-5-5"}
++{"work_id": 4376653350, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.75, "rationale": "Scoping review of computational reviewer-paper matching, a scholarly recommendation and retrieval problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285188099, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Ontology-based meta-learning recommendation of feature selection algorithms; closest to ML methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219176552, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Skill recommendation method using freelancing platform logs; recommender systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221366546, "categories": ["nlp"], "primary": "nlp", "confidence": 0.9, "rationale": "Sentiment analysis of Instagram comments using Naive Bayes and TF-IDF.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135293231, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Information systems thesis on paid Q&A user engagement and social media behavior; closest to social computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204876499, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Digital-library study of scholarly record curatability for open LLM artifacts; the object is metadata and records, not the models.", "model": "claude-sonnet-5-5"}
++{"work_id": 3175543912, "categories": ["information_retrieval", "hci"], "primary": "information_retrieval", "confidence": 0.45, "rationale": "Terminology of exploratory search and information architecture; user information-seeking behavior.", "model": "claude-sonnet-5-5"}
++{"work_id": 4319781162, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.85, "rationale": "Recommender system (collaborative topic regression) with vocabulary completion via a pretrained language model.", "model": "claude-sonnet-5-5"}
++{"work_id": 3208454977, "categories": ["information_retrieval", "nlp"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Dataset recommendation framed as text classification for research problem descriptions; retrieval/recommendation focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391682048, "categories": ["information_retrieval", "privacy_preserving_computation"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Comparative study of recommender algorithms on perturbed ratings (differential privacy).", "model": "claude-sonnet-5-5"}
++{"work_id": 7125905211, "categories": ["information_retrieval", "deep_learning_architectures"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "POI recommendation/CTR prediction with a graph-Mamba model.", "model": "claude-sonnet-5-5"}
++{"work_id": 2888686055, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.75, "rationale": "Weather and location aware recommender system using HMM with IoT sensor data.", "model": "claude-sonnet-5-5"}
++{"work_id": 2798885959, "categories": ["information_retrieval", "hci"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Study of web search result readability and relevance for dyslexic searchers; accessibility in search.", "model": "claude-sonnet-5-5"}
++{"work_id": 3093888250, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.55, "rationale": "Long-tailed sequential user behavior modeling, transferable parameters for recommendation; abstract minimal.", "model": "claude-sonnet-5-5"}
++{"work_id": 7206221668, "categories": [], "primary": null, "confidence": 0.95, "rationale": "University library information bulletin, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2907543384, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "Ontology-based personalized learning path recommendation system.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138937744, "categories": ["information_retrieval", "multimodal_ai"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "E-commerce recommendation model with multimodal fusion, large representation learning and RL-based marketing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3036207666, "categories": ["information_retrieval", "ai_safety"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Fairness framework for recommender and search systems in two-sided marketplaces.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409095591, "categories": ["information_retrieval", "multimodal_ai", "hci"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Knowledge-graph classification of multimodal medical VR resources plus a recommender.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399261954, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Minimal abstract on web scraping for product analysis; no clear category contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4224219747, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Book chapter on latent factors in collaborative filtering recommenders.", "model": "claude-sonnet-5-5"}
++{"work_id": 7172068451, "categories": ["information_retrieval", "software_engineering"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Neural code-to-code retrieval system with FAISS index, evaluated against BM25.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221540348, "categories": ["hci"], "primary": "hci", "confidence": 0.6, "rationale": "Web prototype recognizing children from their interaction (image selection) patterns, tested in a user study.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160413516, "categories": ["information_retrieval", "multimodal_ai", "llm"], "primary": "multimodal_ai", "confidence": 0.6, "rationale": "Vision-language-model visual attribution for iterative RAG over document screenshots.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415775831, "categories": ["llm"], "primary": "llm", "confidence": 0.7, "rationale": "RAG system with QLoRA-tuned Mistral-7B for maritime assessment feedback.", "model": "claude-sonnet-5-5"}
++{"work_id": 4306317060, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.85, "rationale": "Consistency metric for evaluating neural IR ranking models across query reformulations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409365993, "categories": ["ml_theory", "information_retrieval"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Prioritized multi-task optimization with Lagrangian multiplier methods and guarantees, applied to search ranking.", "model": "claude-sonnet-5-5"}
++{"work_id": 2883625840, "categories": ["information_retrieval", "hci"], "primary": "information_retrieval", "confidence": 0.7, "rationale": "Thesis on a tourism destination recommender built with ML classifiers, ensembles and an adaptive user interface.", "model": "claude-sonnet-5-5"}
++{"work_id": 4224311926, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Interpretable collaborative filtering for attribute-based recommendation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131644180, "categories": ["information_retrieval", "llm"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "LLM offline reasoning to build persona-based item indexes for efficient recommendation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404884507, "categories": ["privacy_preserving_computation", "blockchain", "information_retrieval"], "primary": "privacy_preserving_computation", "confidence": 0.6, "rationale": "Combines differential privacy and blockchain for privacy-preserving recommender systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3208034694, "categories": ["software_engineering", "hci"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Automated accessibility testing of dynamic web applications via a crawler framework.", "model": "claude-sonnet-5-5"}
++{"work_id": 4377079764, "categories": ["information_retrieval", "deep_learning_architectures"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Multi-interest sequential recommendation model using capsule networks and attention.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412377349, "categories": ["information_retrieval", "deep_learning_architectures"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Graph debiasing for next POI recommendation, a recommender systems task using GNNs.", "model": "claude-sonnet-5-5"}
++{"work_id": 3094673112, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.35, "rationale": "Survey of AI systems for tourism including recommender systems; only loosely a recommendation focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 7206653269, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Library newsletter, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2801495198, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Short book review on internet searching in a librarianship journal, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3000081209, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.9, "rationale": "Collaborative filtering movie recommender implementation and evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128791697, "categories": ["information_retrieval", "computer_vision", "hci"], "primary": "information_retrieval", "confidence": 0.6, "rationale": "Emotion-aware video recommender using CNN facial emotion detection and GA-PSO optimisation; affective computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389683448, "categories": ["hci"], "primary": "hci", "confidence": 0.55, "rationale": "Survey-based study of how recommender quality affects consumer purchase intention; user-behaviour study.", "model": "claude-sonnet-5-5"}
++{"work_id": 6910675515, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Bibliometric/sociological study of academia-industry norms in AI research; not core CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391150058, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.4, "rationale": "Personalization techniques for semantic web, search and agent synthesis; loosely web/IR.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220856794, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Theological and narrative identity literature review about recommendation algorithms; not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7217759194, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.8, "rationale": "Category-based personalized search using session clustering with SVM.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160261118, "categories": ["ml_theory", "deep_learning_architectures"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Neural causal inference for individual treatment effect estimation under imbalance and cold-start.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415702047, "categories": ["privacy_preserving_computation", "information_retrieval"], "primary": "privacy_preserving_computation", "confidence": 0.8, "rationale": "Defends recommender embeddings against attribute inference attacks via gradient obfuscation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160178876, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.5, "rationale": "Website structure analysis and automatic hyperlink reconfiguration using page similarity, for navigation and search indexing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3169195854, "categories": ["compilers"], "primary": "compilers", "confidence": 0.8, "rationale": "Studies loop transformations and polyhedral model for compiler optimization performance.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412395283, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Clinical randomized trial of a mHealth app; not CS research focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 2798370950, "categories": ["compilers"], "primary": "compilers", "confidence": 0.92, "rationale": "JIT compiler design for PHP/Hack at PLDI.", "model": "claude-sonnet-5-5"}
++{"work_id": 3179712771, "categories": [], "primary": null, "confidence": 0.8, "rationale": "ESL error analysis and a grammar website; language teaching, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213884990, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Community service literacy training for high school students; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 1555598009, "categories": ["compilers"], "primary": "compilers", "confidence": 0.9, "rationale": "Code generator with stack scheduling for stack-based execution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221269676, "categories": [], "primary": null, "confidence": 0.6, "rationale": "AutoLISP script for electrical design automation; engineering application, little CS research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 2591985246, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Curriculum reform in computer organization education; not a CS research topic in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3109527096, "categories": ["compilers"], "primary": "compilers", "confidence": 0.93, "rationale": "Compiler optimization with online profiling evaluated on LLVM.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407848518, "categories": ["compilers"], "primary": "compilers", "confidence": 0.93, "rationale": "MLIR Transform dialect for controlling compiler transformations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4282571853, "categories": [], "primary": null, "confidence": 0.6, "rationale": "DC motor speed control with PID and fuzzy logic is control engineering, not covered by any computer-science category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410182237, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Educational puzzle game using a forward chaining rule. It is a simple application with no clear research contribution in any category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4321500346, "categories": ["compilers", "programming_languages"], "primary": "compilers", "confidence": 0.8, "rationale": "CARMOT is a compiler and runtime tool for characterizing program state, supporting OpenMP and smart pointer abstractions.", "model": "claude-sonnet-5-5"}
++{"work_id": 2908836625, "categories": ["compilers"], "primary": "compilers", "confidence": 0.3, "rationale": "The title concerns optimizing the compilation of probabilistic business rules using PRM, but the abstract is empty, so the rating rests on the title alone.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128400621, "categories": ["compilers", "quantum_computing", "programming_languages"], "primary": "compilers", "confidence": 0.88, "rationale": "Adapts classical compiler passes (static analyses) to quantum programs, which also touches quantum software and program analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386528956, "categories": ["compilers", "ai_accelerators", "distributed_computing"], "primary": "compilers", "confidence": 0.7, "rationale": "Direct GPU compilation through LLVM/OpenMP offloading, using ensemble execution to improve GPU parallelism and utilization.", "model": "claude-sonnet-5-5"}
++{"work_id": 2901311712, "categories": ["compilers", "ai_accelerators"], "primary": "compilers", "confidence": 0.6, "rationale": "CUDA kernel block size tuning and mixed precision tuning, using performance models and search for GPU program optimization.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401497158, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "CMOS circuit design of an inductor-current sensor for a DC-DC converter. This is analog/VLSI hardware design and only loosely fits computer architecture.", "model": "claude-sonnet-5-5"}
++{"work_id": 3188260334, "categories": ["compilers", "programming_languages"], "primary": "compilers", "confidence": 0.75, "rationale": "Managed-runtime optimization (lookup caches, monomorphization bottlenecks) based on program execution phases, which is JIT and VM compilation work.", "model": "claude-sonnet-5-5"}
++{"work_id": 2992990171, "categories": [], "primary": null, "confidence": 0.95, "rationale": "College English teaching methods are education research, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415934317, "categories": ["compilers", "programming_languages"], "primary": "compilers", "confidence": 0.8, "rationale": "Program specialization via transformations and a Cloud Sisal compiler for parallel code.", "model": "claude-sonnet-5-5"}
++{"work_id": 3020776133, "categories": ["ai_accelerators", "compilers"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Translator from NNEF IR to clDNN to speed up NN inference on OpenCL/Intel devices.", "model": "claude-sonnet-5-5"}
++{"work_id": 2794424394, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Control theory for memristive neural networks; not within the CS categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206295199, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Relevance vector machine (Bayesian method) for electrode selection in tomography imaging.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403183224, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Control-theoretic stability analysis of semi-Markov jump systems; not CS in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4380434351, "categories": ["compilers", "computer_architecture"], "primary": "compilers", "confidence": 0.8, "rationale": "Compilation and optimisation for heterogeneous embedded multicore platforms, source-to-source compilation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3134130217, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Educational study of digital resources in early education, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2906120018, "categories": ["programming_languages", "software_engineering"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Static code analyzer using SSA and dataflow analysis for an educational algorithms portal.", "model": "claude-sonnet-5-5"}
++{"work_id": 2897836007, "categories": ["compilers"], "primary": "compilers", "confidence": 0.85, "rationale": "Automatic loop parallelization by detecting scans and reductions in recurrences.", "model": "claude-sonnet-5-5"}
++{"work_id": 3193709667, "categories": ["compilers", "algorithms"], "primary": "compilers", "confidence": 0.8, "rationale": "Polyhedral-style parallel tiled code generation for linear recurrences with OpenMP.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392877469, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Fuzzy control implemented on an AT89C51 microcontroller-based process controller; embedded hardware/software, weak fit to the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 2981857535, "categories": ["database_systems", "compilers"], "primary": "database_systems", "confidence": 0.8, "rationale": "Machine code caching for the PostgreSQL query JIT compiler, covering both query execution and JIT compilation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2886921199, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Iterative learning control for wind turbines is control engineering, not matching any CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4226087031, "categories": ["edge_computing", "hci"], "primary": "edge_computing", "confidence": 0.4, "rationale": "Smart-city sensor network instrument deployment with community engagement; IoT platform as computing substrate.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392445319, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Sequence mining of student game logs to understand problem-solving behavior in a learning environment.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416730187, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Development of a teacher administration application with the ADDIE model and validation; applied software development.", "model": "claude-sonnet-5-5"}
++{"work_id": 3138455952, "categories": ["compilers", "database_systems", "ai_accelerators"], "primary": "compilers", "confidence": 0.8, "rationale": "Query compilation on GPUs: code generation and optimization for database queries on GPU hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 2912097352, "categories": ["compilers", "applied_cryptography"], "primary": "compilers", "confidence": 0.75, "rationale": "Compilation technique that works under hardware aliasing and encrypted computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4384026566, "categories": ["compilers"], "primary": "compilers", "confidence": 0.85, "rationale": "Automatic optimizer for straightline code, outperforming off-the-shelf compilers; code generation and optimization.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411259674, "categories": ["compilers", "formal_verification"], "primary": "compilers", "confidence": 0.75, "rationale": "Self-adaptive compilation framework switching between verified and optimized compilation for embedded systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 7142973430, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Education assessment instrument development; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3161777593, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Vocational English textbook reform; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4290851516, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.75, "rationale": "Domain adaptation source selection for Sentinel-2 remote sensing imagery regression.", "model": "claude-sonnet-5-5"}
++{"work_id": 7203891956, "categories": ["programming_languages", "compilers"], "primary": "programming_languages", "confidence": 0.8, "rationale": "Region-based memory management and tail recursion semantics; relates to continuation marks in language implementation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3082269642, "categories": ["compilers", "computer_architecture"], "primary": "compilers", "confidence": 0.7, "rationale": "Automated transformation of Simulink models via IBSDF intermediate representation to multi-core embedded code; compile-time verification of properties.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386632677, "categories": ["programming_languages", "compilers"], "primary": "compilers", "confidence": 0.65, "rationale": "Extensible compact Scheme VM where the compiler generates specialized VMs; VM implementation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160557760, "categories": ["compilers", "ai_accelerators"], "primary": "compilers", "confidence": 0.85, "rationale": "AOT compiler for PyTorch CPU inference emitting LLVM IR; ML compiler.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131081605, "categories": ["compilers", "database_systems"], "primary": "compilers", "confidence": 0.7, "rationale": "Source-to-source DataFrame query optimizer with IR and rewrite rules, published at CGO.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403493438, "categories": ["compilers", "programming_languages"], "primary": "compilers", "confidence": 0.85, "rationale": "Generating abstract-interpretation optimizer for CPython JIT from a DSL.", "model": "claude-sonnet-5-5"}
++{"work_id": 7115276129, "categories": ["compilers", "programming_languages"], "primary": "compilers", "confidence": 0.6, "rationale": "Common compiler ecosystem for DSLs; minimal abstract but title is clear.", "model": "claude-sonnet-5-5"}
++{"work_id": 2905512450, "categories": ["speech_audio", "hci"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Audio spatialisation library with hearing loss and hearing aid simulation for virtual environments; audio processing with an HCI/accessibility angle.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415312216, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Learning-based sound field reproduction for automotive audio systems is audio signal processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219107916, "categories": ["speech_audio", "hci"], "primary": "speech_audio", "confidence": 0.75, "rationale": "ResNet-50 environmental sound classification in an Android app for hearing-impaired users, with a usability evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2805996599, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Age identification from voice using GMM supervectors and DNNs is speaker characterisation, i.e. speech processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388462287, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.5, "rationale": "Book chapter on microphone use techniques in audio recording; a practical audio engineering topic with weak computer-science content.", "model": "claude-sonnet-5-5"}
++{"work_id": 4307835756, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Sparse array design for DOA estimation is array signal processing for radar and sonar, which fits none of the computer-science categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414880351, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.75, "rationale": "Acoustic drone detection using Hilbert-Huang Transform, MFCC features and a lightweight neural network is audio event detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4384303965, "categories": ["speech_audio", "multimodal_ai"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Audio-visual multi-channel speech separation, dereverberation and recognition; mainly speech processing with a visual modality.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385474051, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Data augmentation and pre-training for end-to-end ASR.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386436058, "categories": ["multimodal_ai", "speech_audio"], "primary": "multimodal_ai", "confidence": 0.75, "rationale": "Audio-video contrastive dual-encoder learning for music videos, evaluated on music tagging and genre classification.", "model": "claude-sonnet-5-5"}
++{"work_id": 6966747421, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "MFCC-based analysis of L1 influence on L2 pronunciation; acoustic speech analysis with classical ML.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162373303, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Music source separation comparing DenseNet-based models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387489607, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.35, "rationale": "Deep learning DOA estimation for array signals; signal processing with autoencoder and VGG networks, no clean category fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392902829, "categories": ["speech_audio", "multimodal_ai"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Challenge overview for audio-visual target speaker extraction, benchmark for speech front-end.", "model": "claude-sonnet-5-5"}
++{"work_id": 3156731511, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Language model integration for attention encoder-decoder ASR.", "model": "claude-sonnet-5-5"}
++{"work_id": 7139231366, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Harp performance dataset and DTW-based rhythm analysis; music information retrieval.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155062195, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Speech restoration using latent acoustic features and diffusion, fast inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 2936511587, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.92, "rationale": "Speaker recognition embeddings and domain adaptation with autoencoders.", "model": "claude-sonnet-5-5"}
++{"work_id": 2910678517, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Geometric calibration of distributed microphone arrays; acoustic signal processing and source localisation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412729123, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Music mixing practice book chapter; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400037432, "categories": ["speech_audio", "hci"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Speech-based emotion recognition using bispectrum images and a CNN; affective computing is a secondary fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 2883033779, "categories": ["speech_audio", "algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Blind source separation solved with an Invasive Weed Optimization metaheuristic; the contribution is mainly the optimization algorithm, with signal separation as the application.", "model": "claude-sonnet-5-5"}
++{"work_id": 2991471437, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Music information retrieval: HMM-based structure analysis of sonata form with a released annotated dataset.", "model": "claude-sonnet-5-5"}
++{"work_id": 4327772893, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Deep learning for music score recognition and algorithmic composition; falls under music computing and generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7172368152, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Physical acoustics and mechanics of piano performance; it is not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166639495, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.92, "rationale": "Multi-channel target speaker extraction for ASR in multi-party conversations, with no diarization step.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169835704, "categories": ["generative_ai", "speech_audio"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Zero-shot music editing with a rectified-flow diffusion transformer and a high-order ODE solver; this is generative modelling applied to audio.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411133045, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Multi-label bird species classification from audio using CNN transfer learning and an LSTM; this is audio event detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140139406, "categories": ["speech_audio", "nlp"], "primary": "speech_audio", "confidence": 0.88, "rationale": "Irish ASR transcription system with semi-supervised acoustic modelling, plus punctuation and capitalisation restoration.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411733613, "categories": ["hci", "speech_audio"], "primary": "hci", "confidence": 0.6, "rationale": "Empirical user study of feedback conditions in an ASR-based reading tutor for children; the contribution is the educational interaction study, not the ASR itself.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405633169, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Continual learning for ASR using a machine speech chain with TTS replay; speech recognition is the core.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313196217, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.88, "rationale": "CNN-based classification of Vietnamese folk songs is music information retrieval.", "model": "claude-sonnet-5-5"}
++{"work_id": 3197227964, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Personalized speech enhancement using self-supervised data purification.", "model": "claude-sonnet-5-5"}
++{"work_id": 2932055495, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.85, "rationale": "End-to-end CNN binaural sound source localisation from raw waveform; audio processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7122349630, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Systematic survey of automatic music plagiarism detection, which is music information retrieval.", "model": "claude-sonnet-5-5"}
++{"work_id": 2980513516, "categories": ["speech_audio", "cybersecurity"], "primary": "speech_audio", "confidence": 0.8, "rationale": "Spoofing countermeasures for speaker verification using glottal features; speech processing with a security angle.", "model": "claude-sonnet-5-5"}
++{"work_id": 2965346518, "categories": ["hci"], "primary": "hci", "confidence": 0.45, "rationale": "Musical score representation using colour and shape, based on cross-modal mapping; visual notation design is closest to HCI.", "model": "claude-sonnet-5-5"}
++{"work_id": 3094199623, "categories": ["speech_audio", "deep_learning_architectures"], "primary": "speech_audio", "confidence": 0.88, "rationale": "Real-time speech enhancement using a dual-path self-attention RNN.", "model": "claude-sonnet-5-5"}
++{"work_id": 3196692235, "categories": ["speech_audio", "computational_biology"], "primary": "speech_audio", "confidence": 0.7, "rationale": "COVID-19 detection from cough sounds using acoustic features; audio analysis with a health application.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386556071, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.75, "rationale": "VAE-based disentanglement of harmonic and rhythmic features in music audio, applied to remix generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390905911, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.93, "rationale": "Streaming ASR domain adaptation with CTC decoding is core speech recognition.", "model": "claude-sonnet-5-5"}
++{"work_id": 2982577960, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.6, "rationale": "Single-channel blind source separation of mixed signals including speech; audio separation focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 3160275121, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.9, "rationale": "Multilingual phonetic dataset for low-resource phone recognition.", "model": "claude-sonnet-5-5"}
++{"work_id": 4212954326, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Underwater array beamforming signal processing; not clearly a CS category in this ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3134129880, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Physics World popular article on smart speaker acoustic sensors; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7156277776, "categories": ["speech_audio", "edge_computing"], "primary": "speech_audio", "confidence": 0.7, "rationale": "Acoustic drone detection and localization with a compact CNN, deployed as TinyML on embedded hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148385849, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Evaluates self-supervised speech representations for diarization and separation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3033627755, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.93, "rationale": "End-to-end neural speaker diarization method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415338423, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.92, "rationale": "Benchmarking of speaker diarization models across datasets.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411173190, "categories": ["speech_audio"], "primary": "speech_audio", "confidence": 0.7, "rationale": "sEMG-based unvoiced speech word classification with ML models.", "model": "claude-sonnet-5-5"}
++{"work_id": 7105655369, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Scoping review protocol on knowledge graphs and semantic technologies for data integration in sport science; closest fit is data management, weakly.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412742458, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Systematic review of AI-generated instructional videos (AIGC pipeline: image generation, speech synthesis, animation) and their learning impacts.", "model": "claude-sonnet-5-5"}
++{"work_id": 3097725556, "categories": ["database_systems", "software_engineering"], "primary": "database_systems", "confidence": 0.55, "rationale": "Process mining connector extracting event logs from the Camunda workflow engine; data extraction and mining of process data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4317502683, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Data integration of hospital and management information systems and its effect on report accuracy.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411208804, "categories": ["cloud_computing", "software_engineering", "distributed_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Scalable architecture patterns (microservices, CQRS, Kubernetes, serverless) for AI-CRM systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 7107958642, "categories": ["database_systems", "programming_languages"], "primary": "database_systems", "confidence": 0.65, "rationale": "Course on Datalog, its extensions and Datalog engines for computing on structured data.", "model": "claude-sonnet-5-5"}
++{"work_id": 2991196924, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Only a title is available, suggesting integrating linked open data into a data cube. The abstract is empty, so there is too little information to assign a category confidently.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414507907, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.6, "rationale": "Hybrid SQL/NoSQL databases with AI-driven optimizations; only the title is informative.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225908445, "categories": ["database_systems", "algorithms"], "primary": "database_systems", "confidence": 0.85, "rationale": "Efficient skyline query processing algorithm on massive incomplete data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4307637929, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Survey of managers on human-AI collaboration in strategic marketing decisions; weak CS content, closest to human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 7218431496, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Applied data mining with C4.5 classification for sales prediction; closest category is data mining as data management, though weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 3018396859, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.8, "rationale": "Experimental study of Hive query performance on SQL-on-Hadoop engines and data distribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 2604259521, "categories": ["database_systems", "nlp"], "primary": "database_systems", "confidence": 0.65, "rationale": "Knowledge base construction system from richly formatted data with a data model and deep learning extraction; SIGMOD data management.", "model": "claude-sonnet-5-5"}
++{"work_id": 2920557606, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.4, "rationale": "Performance analysis of hyperspectral iris imaging instrument; image acquisition system evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2951532316, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.85, "rationale": "Database extension for spatio-temporal data streams with continuous queries and window operators.", "model": "claude-sonnet-5-5"}
++{"work_id": 4327772251, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Generic data mining overview in information management systems; weakly data mining as data management.", "model": "claude-sonnet-5-5"}
++{"work_id": 2799038541, "categories": ["database_systems", "algorithms"], "primary": "database_systems", "confidence": 0.7, "rationale": "Index structure for MSO query enumeration under updates; PODS database theory with algorithmic flavor.", "model": "claude-sonnet-5-5"}
++{"work_id": 4289533871, "categories": ["database_systems", "algorithms", "distributed_computing"], "primary": "database_systems", "confidence": 0.8, "rationale": "Distributed subgraph matching framework implemented on Pregel+, graph query processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2942479669, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.8, "rationale": "Converting unstructured data to structured form using NoSQL Couchbase document database.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131351419, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "AI-assisted transaction processing system for counselors; applied information system with effectiveness evaluation, weak fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 7159658581, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Applies a Naive Bayes classifier to hiring decisions; a routine application with no clear contribution to any ontology category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401268941, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Library science textbook on data librarianship; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161581487, "categories": ["database_systems", "information_retrieval"], "primary": "database_systems", "confidence": 0.85, "rationale": "Filtered vector ANNS indexing with selectivity estimation in a data management venue (PACMMOD); vector database query processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387233410, "categories": [], "primary": null, "confidence": 0.45, "rationale": "Routine application of a neural network with PSO to predict student graduation; no clear contribution to a specific category.", "model": "claude-sonnet-5-5"}
++{"work_id": 2997642018, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Development of a payroll information system using the prototype method, PHP and MySQL; an applied software development project.", "model": "claude-sonnet-5-5"}
++{"work_id": 3168932797, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.7, "rationale": "Record linkage (deterministic and probabilistic) for data integration across health databases; data integration and cleaning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118687013, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.5, "rationale": "Systematic review of generative AI and deep learning in personalized education; covers adaptive content creation and human-AI aspects.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401822397, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Routine application of KNN to predict student graduation; no clear contribution to a specific category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3134393417, "categories": ["software_engineering", "database_systems"], "primary": "software_engineering", "confidence": 0.75, "rationale": "Identifies faults in data pipelines and mitigation strategies, validated with action research in a software organization at APSEC.", "model": "claude-sonnet-5-5"}
++{"work_id": 3201407754, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Routine application of standard data mining methods to predict vocational students' job-market readiness; no clear contribution to a specific category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4205179238, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Development of a library loan information system using R&D method, Java and MySQL; an applied system-building paper.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416374389, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Web system for sales and inventory processes built with PHP/MySQL and evaluated with a questionnaire.", "model": "claude-sonnet-5-5"}
++{"work_id": 4255118578, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.6, "rationale": "SQL Server wait statistics baselining for performance monitoring.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160618846, "categories": ["llm", "database_systems"], "primary": "llm", "confidence": 0.5, "rationale": "Observability platform for LLM applications with trace ingestion and hybrid time-series/vector storage, benchmarked.", "model": "claude-sonnet-5-5"}
++{"work_id": 2982402713, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.95, "rationale": "Garbage collection for MVCC in HTAP database systems; transaction processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 6944527151, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.9, "rationale": "Analytical database system Actian Vector: query planning, integrations, usability.", "model": "claude-sonnet-5-5"}
++{"work_id": 3036960336, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.4, "rationale": "Health informatics dashboard for influenza surveillance data quality feedback in primary care.", "model": "claude-sonnet-5-5"}
++{"work_id": 2933476764, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.6, "rationale": "Platform replicating data transparently between different DBMSs, a software-defined database cluster.", "model": "claude-sonnet-5-5"}
++{"work_id": 3008451085, "categories": ["database_systems", "distributed_computing"], "primary": "database_systems", "confidence": 0.9, "rationale": "Highly available HTAP DBMS with a two-step commit and consensus-based serialization.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402761012, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.5, "rationale": "Portal integrating and searching paleontological specimen records from many museum databases; data integration.", "model": "claude-sonnet-5-5"}
++{"work_id": 7116397644, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.4, "rationale": "Applies K-Means data mining to drug inventory data; weak fit to data mining as a data-management problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 3189460411, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Management research on firm coopetition networks, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3022939691, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.35, "rationale": "Latent representation of event streams with matrix factorization and autoencoders plus forecasting; generic ML methodology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3010181845, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.3, "rationale": "Educational data mining comparing Naive Bayes and KNN classifiers; the only fit is a loose data-mining one.", "model": "claude-sonnet-5-5"}
++{"work_id": 4282592169, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.95, "rationale": "Automatic schema normalization learning for denormalized tables, published at SIGMOD.", "model": "claude-sonnet-5-5"}
++{"work_id": 3043717349, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Belief representations (Dempster-Shafer, fuzzy, probabilistic) and their interoperable conversion; probabilistic modeling.", "model": "claude-sonnet-5-5"}
++{"work_id": 7159820960, "categories": ["database_systems", "cloud_computing"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Review of integrating relational databases with cloud environments.", "model": "claude-sonnet-5-5"}
++{"work_id": 4220806123, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.6, "rationale": "Improved Apriori association rule mining with a three-dimensional matrix for efficiency.", "model": "claude-sonnet-5-5"}
++{"work_id": 3183141872, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.55, "rationale": "Frequent itemset and Apriori association rule mining on purchase data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7212066750, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Development of a web-based fitness management system with usability evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4293812983, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.9, "rationale": "Gradual dependent type theory with metatheory: type safety, canonicity, gradual guarantees.", "model": "claude-sonnet-5-5"}
++{"work_id": 3046578987, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Process algebra with Petri net semantics, bisimulation equivalence, and performance analysis of concurrent systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394658783, "categories": ["programming_languages", "computational_complexity"], "primary": "programming_languages", "confidence": 0.45, "rationale": "Categorical semantics of causal morphisms, traced monoidal categories and recursion; theoretical semantics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128354711, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.6, "rationale": "Development of a web inventory information system using the Waterfall method and black box testing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4296940370, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.4, "rationale": "Book chapter wrapping up a Haskell text covering functional programming and type systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3091828950, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Didactics of programming based on Piaget's theory is education research, not a listed CS category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124299435, "categories": ["ai_agents", "formal_verification"], "primary": "ai_agents", "confidence": 0.7, "rationale": "Enactability of multi-agent interaction protocols using trace expressions; multi-agent systems with formal analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4320558734, "categories": ["compilers", "software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "VSCode language support tool (syntax highlighting, autocompletion) for the COOL teaching language.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414483419, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Errata on denotational models of mixed choice in concurrency, using topological and domain-theoretic semantics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128043521, "categories": ["ml_theory", "programming_languages"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Proof-theoretic account of Bayesian inference using linear logic proof-nets and Bayesian networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392822342, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.8, "rationale": "Proves mergesort correctness and stability using parametricity in Rocq.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367318871, "categories": ["programming_languages", "compilers"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Compares WebAssembly and JavaScript execution performance.", "model": "claude-sonnet-5-5"}
++{"work_id": 3194305803, "categories": ["software_engineering", "programming_languages"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Property-based testing library and DSL for stateful models.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206497465, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.5, "rationale": "Book chapter on Rust ownership, moves, and copies.", "model": "claude-sonnet-5-5"}
++{"work_id": 4320928251, "categories": ["software_engineering", "programming_languages", "formal_verification"], "primary": "software_engineering", "confidence": 0.55, "rationale": "Contract-based design with dependent types for extra-functional properties.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164912132, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Synthesizes best abstract transformers using OMT solving for static analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 2969568901, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.92, "rationale": "Type system with linear and non-linear recursive types and categorical semantics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7140223191, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Canonical representation of expression DAGs for symbolic regression search.", "model": "claude-sonnet-5-5"}
++{"work_id": 3204449783, "categories": [], "primary": null, "confidence": 0.8, "rationale": "E-learning awareness survey in education, not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220384869, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Web e-commerce system development and testing for a small business.", "model": "claude-sonnet-5-5"}
++{"work_id": 2955161317, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.6, "rationale": "Recognizable tree languages and closure operations are automata and formal language theory, which this ontology places under computability and automata theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4307886855, "categories": ["programming_languages", "software_engineering"], "primary": "programming_languages", "confidence": 0.85, "rationale": "Derives incremental type-checkers from type system specifications written in a meta-DSL, using scope graphs for name binding.", "model": "claude-sonnet-5-5"}
++{"work_id": 2892794698, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Builds a course information system with the Waterfall method. It is a development case study with little research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413180303, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Power grid flow visualization is mainly electrical engineering, not clearly a computer science contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 6891551213, "categories": ["programming_languages"], "primary": "programming_languages", "confidence": 0.7, "rationale": "Artifact for a paper on shape analysis with gradual types, which is a type systems topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990464075, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Plenary talk abstract on Industries 5.0 and parallel industrial systems. It is not a specific research contribution that fits a category.", "model": "claude-sonnet-5-5"}
++{"work_id": 3199165787, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Development of a web-based alumni information system with structured diagrams and MySQL. It is a routine application build.", "model": "claude-sonnet-5-5"}
++{"work_id": 3010153186, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.45, "rationale": "Variational Bayes nonlinear filter for state estimation. It is a probabilistic inference method.", "model": "claude-sonnet-5-5"}
++{"work_id": 3030633439, "categories": [], "primary": null, "confidence": 0.5, "rationale": "A general discussion of big data in education, with no clear computer science research contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 2941342584, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.3, "rationale": "Development of an Android academic information system. It is an application build with functional testing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215646023, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Community service dissemination of an orchid disease expert system app; application deployment/training rather than clear CS research contribution in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4287283171, "categories": ["formal_verification", "programming_languages"], "primary": "formal_verification", "confidence": 0.55, "rationale": "Dynamic epistemic separation logic with semantics and elimination of modalities; logic-focused, close to program-verification logics.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388795790, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Training of high school students in App Inventor; educational outreach, not core CS research in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386211482, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Logic programming proceedings with abstract interpretation domain design for logic programs; static analysis as language problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 6910567405, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.7, "rationale": "Minimal perfect hash function implementation (PtrHash), a data structure/algorithm engineering artifact.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281724929, "categories": ["programming_languages", "computer_architecture"], "primary": "programming_languages", "confidence": 0.4, "rationale": "Implicit state machines for real-time/reactive embedded systems at LCTES; language abstraction for embedded systems, abstract is minimal.", "model": "claude-sonnet-5-5"}
++{"work_id": 3121979527, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.8, "rationale": "Multiparty session types, pi-calculus type system ensuring protocol conformance and deadlock-freedom; type-theoretic concurrency.", "model": "claude-sonnet-5-5"}
++{"work_id": 7153697985, "categories": ["programming_languages", "cybersecurity", "formal_verification"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Information-flow control via security type systems and secure runtime semantics, with Coq proofs; timing channel attacks.", "model": "claude-sonnet-5-5"}
++{"work_id": 2810401719, "categories": ["programming_languages", "formal_verification"], "primary": "programming_languages", "confidence": 0.6, "rationale": "Cubical type theory and homotopy type theory foundations; type theory with informal proofs of higher groupoid structure.", "model": "claude-sonnet-5-5"}
++{"work_id": 2888336873, "categories": ["formal_verification", "computer_architecture", "software_engineering"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Static WCET estimation with abstract interpretation for PLC control code in real-time automation systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3009620545, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.9, "rationale": "Hyperledger Fabric artefact for maintenance documentation, evaluated on throughput and latency.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416148866, "categories": ["applied_cryptography", "blockchain", "privacy_preserving_computation"], "primary": "applied_cryptography", "confidence": 0.85, "rationale": "New threshold signature scheme with provable security, applied to privacy-preserving lending on blockchain.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401513854, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.7, "rationale": "Overview of blockchain applications for supply chain transparency and traceability.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402990505, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.5, "rationale": "Chapter on blockchain from a consumer perspective; a weak fit, but the topic is blockchain applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401788847, "categories": ["blockchain", "edge_computing"], "primary": "blockchain", "confidence": 0.7, "rationale": "Smart gateways collect device data and upload it to the blockchain, covering data credibility in an IoT setting.", "model": "claude-sonnet-5-5"}
++{"work_id": 2977293541, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.5, "rationale": "Introductory practitioner book on blockchain basics and business impact.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127122082, "categories": ["blockchain", "software_engineering"], "primary": "blockchain", "confidence": 0.75, "rationale": "Template-based generation of smart contract code from legal contracts via parsing and AST techniques.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402966691, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Decentralized charity platform built from smart contracts and DeFi liquidity pool protocols.", "model": "claude-sonnet-5-5"}
++{"work_id": 7152727986, "categories": ["blockchain", "cybersecurity"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain applications for healthcare data security, covering access control and record integrity.", "model": "claude-sonnet-5-5"}
++{"work_id": 2989399708, "categories": ["blockchain", "computer_networks"], "primary": "blockchain", "confidence": 0.65, "rationale": "Smart-contract payments for vehicle platooning in a blockchain-enabled Internet of Vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214298050, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Public opinion survey on crypto regulation in criminal justice; social science, not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3088531038, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.75, "rationale": "Ethereum-based prototype for healthcare data integrity and breach prevention.", "model": "claude-sonnet-5-5"}
++{"work_id": 3014880043, "categories": [], "primary": null, "confidence": 0.35, "rationale": "DBSCAN anomaly detection on Bitcoin prices; it is financial data mining, and the blockchain system itself is not the object of study.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400034579, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.8, "rationale": "Blockchain and IPFS e-vault with smart contracts for legal records.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416430007, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Dataspace connector prototype for decentralized, contract-governed data exchange of digital product passports.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400485117, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Framework integrating blockchain and digital twins for energy performance contracts, plus a systematic review.", "model": "claude-sonnet-5-5"}
++{"work_id": 2901943454, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Systematic literature review of blockchain use cases and system designs for cities.", "model": "claude-sonnet-5-5"}
++{"work_id": 3194765017, "categories": [], "primary": null, "confidence": 0.85, "rationale": "News item about an NFT sale, not research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382770016, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Survey-based regression study of blockchain's effect on corporate governance; management research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407826131, "categories": ["blockchain", "deep_learning_architectures"], "primary": "blockchain", "confidence": 0.55, "rationale": "Survey on integrating blockchain with neural networks for predictive modeling, covering security and scalability challenges.", "model": "claude-sonnet-5-5"}
++{"work_id": 4245301215, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.4, "rationale": "Business book on decentralization, covering blockchain infrastructure and governance; weakly CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3143290424, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.5, "rationale": "Literature review of blockchain research trends in business.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125399881, "categories": ["blockchain", "generative_ai", "cybersecurity"], "primary": "blockchain", "confidence": 0.8, "rationale": "Blockchain-anchored media authentication framework with smart contract registry to counter deepfakes.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117461924, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.35, "rationale": "Survey/SEM study of consumer behavior with blockchain and AI platforms; marginal CS content.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164945802, "categories": ["federated_learning", "blockchain", "ai_safety"], "primary": "federated_learning", "confidence": 0.85, "rationale": "Federated learning with XAI for fraud detection in non-IID blockchain transactions.", "model": "claude-sonnet-5-5"}
++{"work_id": 3103101544, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Blockchain-based P2P energy market pilot on a permissioned chain.", "model": "claude-sonnet-5-5"}
++{"work_id": 3012297310, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.8, "rationale": "Blockchain data-sharing model with improved consensus for pet medical industry.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417496166, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.4, "rationale": "Empirical study of blockchain adoption and consumer trust in SMEs; management-focused.", "model": "claude-sonnet-5-5"}
++{"work_id": 4327927520, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Blockchain-based hospital health record management system.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411360648, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.45, "rationale": "Bibliometric topic-modeling analysis of blockchain research trends in Korea.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160104539, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.8, "rationale": "Chapter reviewing blockchain applications in supply chains, a blockchain application and its evolution.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394910888, "categories": ["blockchain", "cybersecurity", "privacy_preserving_computation"], "primary": "blockchain", "confidence": 0.75, "rationale": "Survey and framework for security and privacy of blockchain systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160380482, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Ethereum smart contract and IPFS based healthcare EMR and appointment system.", "model": "claude-sonnet-5-5"}
++{"work_id": 3095954801, "categories": ["computer_networks", "cybersecurity", "formal_verification"], "primary": "computer_networks", "confidence": 0.8, "rationale": "Thesis on secure time synchronization protocols (NTP, STS), with formal analysis via ProVerif and Tamarin.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400422781, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.5, "rationale": "Analysis of the crypto-currency market and regulation, with blockchain consensus and security trends; borderline economics.", "model": "claude-sonnet-5-5"}
++{"work_id": 7143345003, "categories": ["blockchain", "edge_computing", "cybersecurity"], "primary": "blockchain", "confidence": 0.7, "rationale": "Convergence of IoT and blockchain for secure data sharing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392408184, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.88, "rationale": "Blockchain platform for copyright traceability built on FISCO BCOS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386500200, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Survey on perception of cryptocurrencies among young users; economics/behavioral finance, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4210686246, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Geopolitical and financial policy discussion of digital currencies; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285316846, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Legal and monetary-policy analysis of crypto-assets in Indonesia; not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3173158936, "categories": ["blockchain", "cybersecurity"], "primary": "blockchain", "confidence": 0.6, "rationale": "Book chapter on blockchain with a cybersecurity maturity model for pharmaceutical supply chain governance.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206450404, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Proposes a blockchain energy trading model with new consensus mechanisms and a fairness mechanism.", "model": "claude-sonnet-5-5"}
++{"work_id": 4366000231, "categories": ["blockchain", "cybersecurity"], "primary": "blockchain", "confidence": 0.6, "rationale": "Applies blockchain to secure e-learners' data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134197021, "categories": ["blockchain", "multimodal_ai"], "primary": "blockchain", "confidence": 0.5, "rationale": "Multimodal NFT market dataset covering images, metadata and trading data; it measures a blockchain ecosystem.", "model": "claude-sonnet-5-5"}
++{"work_id": 4379472050, "categories": ["blockchain", "cloud_computing"], "primary": "blockchain", "confidence": 0.7, "rationale": "Blockchain-based supply chain architecture in a cloud setting, evaluated for sustainability.", "model": "claude-sonnet-5-5"}
++{"work_id": 3198186069, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Editorial introduction to an educational technology journal issue; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4298129476, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.85, "rationale": "Evaluates Ethereum transaction performance and scalability.", "model": "claude-sonnet-5-5"}
++{"work_id": 4315572972, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.6, "rationale": "Blockchain-based system for verifying educational documents.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385250044, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Scientometric analysis of fintech and carbon accounting; not a core CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7123350921, "categories": ["software_engineering", "database_systems"], "primary": "software_engineering", "confidence": 0.55, "rationale": "Centralized ledger architecture with idempotent webhooks and a prototype; it is a system design, not on-chain.", "model": "claude-sonnet-5-5"}
++{"work_id": 3186102448, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Applies cloud computing to e-learning infrastructure in academia.", "model": "claude-sonnet-5-5"}
++{"work_id": 3165257421, "categories": ["sustainable_computing", "cloud_computing"], "primary": "sustainable_computing", "confidence": 0.5, "rationale": "Conceptual model for intelligent temperature monitoring and management in smart data centers (cooling, energy).", "model": "claude-sonnet-5-5"}
++{"work_id": 4296044778, "categories": ["sustainable_computing", "cloud_computing"], "primary": "sustainable_computing", "confidence": 0.8, "rationale": "Taxonomy survey of capacity planning for sustainable cloud data centers.", "model": "claude-sonnet-5-5"}
++{"work_id": 2899546521, "categories": ["cloud_computing", "computer_networks"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Model-based rehoming of VNF service chains in network clouds, covering migration and SLOs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4211066124, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.8, "rationale": "Program file placement control in a hierarchical cloud-edge platform.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407737588, "categories": ["cloud_computing", "hci"], "primary": "hci", "confidence": 0.45, "rationale": "Systematic review of the effect of mobile cloud gaming on player experience and accessibility.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412993992, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Chapter on cloud-native architectures (microservices, containers) for finance in food systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3125267005, "categories": ["cloud_computing", "algorithms"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Bat-algorithm metaheuristic for task allocation and resource management in the cloud.", "model": "claude-sonnet-5-5"}
++{"work_id": 2794672104, "categories": ["sustainable_computing", "cloud_computing", "algorithms"], "primary": "sustainable_computing", "confidence": 0.65, "rationale": "Game-theoretic workload exchange among data centers to use green energy and cut energy costs.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410342350, "categories": ["cloud_computing", "algorithms"], "primary": "cloud_computing", "confidence": 0.7, "rationale": "Optimization and heuristics for profit equalization in cloud federations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409764799, "categories": ["software_engineering", "cloud_computing"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Describes the design and implementation of a web-based interview management system, with security measures and AWS services; mainly a system-building paper.", "model": "claude-sonnet-5-5"}
++{"work_id": 3212132676, "categories": ["algorithms", "cloud_computing"], "primary": "algorithms", "confidence": 0.45, "rationale": "Multi-objective optimisation (NSGA-II) for web service composition under QoS fluctuation; service computing, with an optimisation focus.", "model": "claude-sonnet-5-5"}
++{"work_id": 2980085570, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.55, "rationale": "Evaluates cloud-based group work tools and cloud models for end users; a weak research contribution, but cloud-centred.", "model": "claude-sonnet-5-5"}
++{"work_id": 3010373421, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Interview study of organizational roles in inter-organizational smart service systems engineering; an empirical study of engineering projects.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396789527, "categories": ["cloud_computing", "cybersecurity"], "primary": "cloud_computing", "confidence": 0.55, "rationale": "Literature review of security factors influencing cloud adoption by SMEs; cloud adoption and cloud security.", "model": "claude-sonnet-5-5"}
++{"work_id": 2912717503, "categories": ["cloud_computing", "edge_computing", "distributed_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Survey of decentralised cloud models drawing on mobile cloud, ad hoc and edge computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 809949967, "categories": ["distributed_computing", "software_engineering"], "primary": "distributed_computing", "confidence": 0.55, "rationale": "Prototype system for distributed coordination and workflow configuration over web services, including handheld coordination.", "model": "claude-sonnet-5-5"}
++{"work_id": 7130725155, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Survey of risk assessment, anomaly detection and identity access management methods for cloud security using ML.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400263921, "categories": ["cloud_computing", "cybersecurity"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Simulation of cloud backends for connected vehicles, using observability data for anomaly detection; cloud operations and monitoring.", "model": "claude-sonnet-5-5"}
++{"work_id": 3036390872, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.65, "rationale": "Provenance-aware audit trail framework for assessing SaaS provider efficiency in cloud service delivery.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391093673, "categories": ["cloud_computing", "database_systems"], "primary": "cloud_computing", "confidence": 0.4, "rationale": "Educational big data simulation system with two-level resource scheduling and container-based deployment; weak, system-oriented fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407956345, "categories": [], "primary": null, "confidence": 0.6, "rationale": "AHP-based education quality evaluation index system; mainly education research, not a CS contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 2796225596, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "SLA translation for service composition and resource allocation optimisation, which is cloud/service computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117768918, "categories": ["edge_computing", "distributed_computing"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Empirical study of timeout-based fault detection on edge devices under resource stress.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401791482, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.8, "rationale": "ML-based security framework for protecting public cloud data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7208727987, "categories": ["software_engineering", "cloud_computing"], "primary": "software_engineering", "confidence": 0.75, "rationale": "Microservices architecture reporting system evaluated against a monolith, with load and fault-injection tests; an architecture study.", "model": "claude-sonnet-5-5"}
++{"work_id": 3087979613, "categories": ["edge_computing", "cloud_computing"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Review of the transition from cloud to fog computing, covering features and challenges.", "model": "claude-sonnet-5-5"}
++{"work_id": 2893549539, "categories": ["cloud_computing", "sustainable_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "ML-based, energy-aware resource provisioning for in-memory analytics on cloud platforms.", "model": "claude-sonnet-5-5"}
++{"work_id": 3103744586, "categories": ["cloud_computing", "scientific_computing"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Deploying the LSST science pipelines on AWS and assessing performance, scalability and cost.", "model": "claude-sonnet-5-5"}
++{"work_id": 2967724094, "categories": ["cloud_computing", "database_systems"], "primary": "cloud_computing", "confidence": 0.5, "rationale": "Architecture for real-time VaR/ES risk monitoring on a cloud analytics platform and database.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405846634, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Fuzzy clustering for VM resource allocation in cloud data centers.", "model": "claude-sonnet-5-5"}
++{"work_id": 3043719610, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.9, "rationale": "Performance evaluation of serverless/FaaS versus monolith architectures on Azure.", "model": "claude-sonnet-5-5"}
++{"work_id": 3199597950, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Survey of cloud computing issues, with a focus on security.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413436783, "categories": ["cloud_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.7, "rationale": "Threat analysis and a security model for multi-tenant Kubernetes clusters.", "model": "claude-sonnet-5-5"}
++{"work_id": 2967650664, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.75, "rationale": "Migrating industrial control software to IaaS with real-time latency tests.", "model": "claude-sonnet-5-5"}
++{"work_id": 4248407127, "categories": [], "primary": null, "confidence": 0.7, "rationale": "A PowerShell book chapter with no research content in the abstract.", "model": "claude-sonnet-5-5"}
++{"work_id": 2938024881, "categories": ["sustainable_computing", "algorithms"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Multi-objective negotiation of power profiles for renewable-powered data centers using a genetic algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 4243812454, "categories": ["cloud_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Review of cloud security threats and countermeasures.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400232270, "categories": ["cybersecurity", "cloud_computing"], "primary": "cybersecurity", "confidence": 0.85, "rationale": "Cryptojacking attack vector abusing AWS cloud services.", "model": "claude-sonnet-5-5"}
++{"work_id": 3094723118, "categories": ["cloud_computing", "algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Asymptotically optimal threshold-based load balancing with an adaptive threshold, analyzed through a stochastic model.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407606154, "categories": ["cloud_computing", "cybersecurity"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Survey-based study of monitoring and logging (SIEM, anomaly detection) in cloud environments.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401508590, "categories": ["edge_computing", "cloud_computing"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Zygote container planning and task scheduling for edge native applications; NP-hardness proof and heuristic algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 4389723967, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Cold chain logistics monitoring using SPC and a decision tree; it is an agricultural application with only weak computer-science content and no clear category fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134283222, "categories": ["distributed_computing", "scientific_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Parallelized Kubernetes/Ray pipeline for geospatial ML preprocessing and training, evaluated on Arctic air temperature prediction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4297517292, "categories": ["cloud_computing", "operating_systems"], "primary": "cloud_computing", "confidence": 0.6, "rationale": "Book chapter giving an overview of virtualization and Docker containers.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392175659, "categories": ["cloud_computing", "cybersecurity"], "primary": "cybersecurity", "confidence": 0.6, "rationale": "Overview of cloud security threats and mitigation strategies.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401943441, "categories": ["edge_computing", "cloud_computing"], "primary": "edge_computing", "confidence": 0.85, "rationale": "Empirical study of an edge-to-cloud continuum, covering scheduling, load balancing and AI-based IoT applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 3137445566, "categories": ["sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.8, "rationale": "Comparison of data-centre power consumption between rack and blade servers, aimed at energy savings.", "model": "claude-sonnet-5-5"}
++{"work_id": 2900680085, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.4, "rationale": "LDA topic analysis of virtualization patents, framed around virtualization as a cloud enabler; weak fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403332127, "categories": ["cloud_computing", "sustainable_computing"], "primary": "sustainable_computing", "confidence": 0.7, "rationale": "Energy-efficient geographical load balancer for cloud data centres.", "model": "claude-sonnet-5-5"}
++{"work_id": 4320024362, "categories": ["generative_ai", "scientific_computing"], "primary": "generative_ai", "confidence": 0.85, "rationale": "Proposes a three-player WGAN to synthesize time-domain gravitational wave signals; the contribution is GAN training stability, applied to physics data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118421194, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.93, "rationale": "Noise optimization to mitigate mode collapse and improve diversity in text-to-image diffusion models.", "model": "claude-sonnet-5-5"}
++{"work_id": 7110667317, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Practical guide to using GenAI tools in history teaching; an application and education-oriented use of generative AI, not core CS research. Closest fit is human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 7166546311, "categories": ["hci"], "primary": "hci", "confidence": 0.35, "rationale": "Models the effect of GenAI use on student learning retention with clustering and regression; it studies human-AI use in education rather than advancing ML methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 4379985046, "categories": ["generative_ai", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.85, "rationale": "Text-driven facial image editing in StyleGAN latent space using CLIP; this is image synthesis and editing guided by text.", "model": "claude-sonnet-5-5"}
++{"work_id": 2999632917, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Uses a conditional GAN for image-to-image translation to synthesize multi-frequency PolSAR images; the images are remote sensing data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405955397, "categories": ["generative_ai", "computational_biology"], "primary": "generative_ai", "confidence": 0.6, "rationale": "Systematic review of generative methods (GANs, VAEs, transformers) that decode EEG signals into images, video and audio.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409570677, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Uses a VAE as a data-driven parametrization of CMB power spectra for cosmology; this is ML inside a scientific pipeline.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413402631, "categories": ["hci"], "primary": "hci", "confidence": 0.5, "rationale": "Qualitative narrative inquiry into how international students use generative AI in education; a user study of human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381736219, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.93, "rationale": "Proposes a conditional Wasserstein DCGAN with theoretical justification, aimed at stable training and controllable output.", "model": "claude-sonnet-5-5"}
++{"work_id": 7155650775, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Conceptual, education-focused paper on human-AI mediation in learning; closest to human-AI interaction, though not strongly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7152054860, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.95, "rationale": "Thesis on controlling GANs and VAEs and on scene generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414361486, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Qualitative education-management study of course coordinators; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164539746, "categories": ["generative_ai", "computer_graphics", "multimodal_ai"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Text- and speech-driven 3D talking-head synthesis with SDS text-to-3D and a differentiable renderer.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165127855, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Education and humanities study of AI adoption among history students; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4404404823, "categories": ["generative_ai", "ml_theory"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Diffusion bridge model for time series forecasting, a generative modelling contribution with a theoretical framework.", "model": "claude-sonnet-5-5"}
++{"work_id": 7220858792, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Questionnaire study of curriculum design in a university program; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406954321, "categories": ["generative_ai", "ml_theory"], "primary": "generative_ai", "confidence": 0.88, "rationale": "Momentum Schr\u00f6dinger bridge diffusion method with convergence proofs and image and time-series generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204970079, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Learning analytics of GenAI virtual patient dialogues in medical education; weakly CS, closest to human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4381744251, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Unsupervised domain adaptation for image recognition using GANs and knowledge distillation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402504098, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Improved CycleGAN to augment metal wear image datasets; GAN-based image synthesis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3015491268, "categories": ["generative_ai", "ml_theory"], "primary": "generative_ai", "confidence": 0.75, "rationale": "Conditional GAN with mixture density output for probabilistic time-series forecasting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400444727, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.88, "rationale": "Proposes evaluation protocol and studies metrics (FID, IS) for generative models.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127452594, "categories": ["speech_audio", "generative_ai"], "primary": "speech_audio", "confidence": 0.85, "rationale": "Audio forgery localization with mixture of experts; detection of AI-generated partial audio.", "model": "claude-sonnet-5-5"}
++{"work_id": 7113405664, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Supplementary materials only with no abstract content; insufficient information to classify.", "model": "claude-sonnet-5-5"}
++{"work_id": 2966462029, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Community consistency measurement in dynamic networks using a generative model; graph mining method.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167727964, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Survey of dance students' perceptions of GenAI in education; not core computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2810722879, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.88, "rationale": "Extends ambient modules to GAN hidden space for training on noisy data.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125473265, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Platform studies and moral economy of recommender systems; a social science analysis, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135177077, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education study integrating GenAI into an engineering course; pedagogy rather than CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200011910, "categories": ["computer_vision", "generative_ai"], "primary": "computer_vision", "confidence": 0.7, "rationale": "Super-resolution GAN architecture improvement (ESRGAN variant) is image restoration; GAN-based.", "model": "claude-sonnet-5-5"}
++{"work_id": 7215117915, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Qualitative communication study of Instagram posts; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4395661446, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.55, "rationale": "Remote sensing SAR wave spectra retrieval with CGAN and XGBoost; applied remote-sensing analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3130253613, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Review of GAN-based image translation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171857204, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Education psychology questionnaire study of teachers; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392008307, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.4, "rationale": "Comparison guide of generative AI tools for new users; weak research contribution but about GenAI tools.", "model": "claude-sonnet-5-5"}
++{"work_id": 7143415918, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Business management application study using secondary data; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407424148, "categories": ["generative_ai", "multimodal_ai", "ml_theory"], "primary": "generative_ai", "confidence": 0.7, "rationale": "Framework using diffusion-generated synthetic data for multimodal prediction with statistical guarantees.", "model": "claude-sonnet-5-5"}
++{"work_id": 7112438470, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Education report on GenAI in architecture education; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125351784, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.8, "rationale": "Diffusion-parameterized MoE neural CDE for time series generation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406930504, "categories": ["generative_ai", "computational_biology"], "primary": "generative_ai", "confidence": 0.85, "rationale": "Training-free guidance of diffusion models for geometry-conditioned molecular generation in drug design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412083660, "categories": ["generative_ai", "hci"], "primary": "generative_ai", "confidence": 0.55, "rationale": "Generates and evaluates AI images of festivals with a public survey on uncanny valley perception.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387666097, "categories": ["computer_vision"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Local color transfer between images using segmentation and optimal transport; image processing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7118194253, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Educational study of AI literacy in beauty design learners, not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127651506, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.3, "rationale": "Application of AIGC to cultural visual design; weak CS contribution, mostly a design workflow.", "model": "claude-sonnet-5-5"}
++{"work_id": 4367292510, "categories": ["generative_ai", "deep_learning_architectures"], "primary": "generative_ai", "confidence": 0.8, "rationale": "GAN-based generative model for ship hull geometry design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2904629716, "categories": ["computer_vision", "generative_ai"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Attentional GAN for large-scale single-image super resolution, with distributed training; image restoration.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406160051, "categories": ["generative_ai", "computer_vision"], "primary": "generative_ai", "confidence": 0.4, "rationale": "Magazine piece on using generative AI hallucinations for scientific image deblurring.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138839474, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.8, "rationale": "Sketch recognition using improved CycleGAN features and dual attention classifier.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133211554, "categories": ["ml_theory", "algorithms"], "primary": "ml_theory", "confidence": 0.7, "rationale": "New sliced optimal transport distance approximation with projection technique, evaluated on color transfer.", "model": "claude-sonnet-5-5"}
++{"work_id": 7116068851, "categories": ["ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Trains interpretability assistants that read model activations to detect jailbreaks and implanted concepts. This is interpretability work on LLMs, so it fits ai_safety.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417167881, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Management and governance study applying the technology acceptance model and actor-network theory. It is not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2910154596, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.55, "rationale": "Applies the software engineering code of ethics to Tesla Autopilot. It is software engineering ethics, though not strictly a technical contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154958542, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Applies BERT-based text analysis to app reviews of an AI companion. The method is text mining, though the framing is consumer behaviour.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133246977, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Survey of student attitudes toward AI in higher education. This is education and social science research, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7119528090, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Legal analysis of liability for generative AI providers when models produce CSAM. It is a governance and regulation topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 7168242482, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Landscape analysis of ethical perspectives on generative AI in computing education. It belongs to education research, with no clear fit in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 7151628917, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.9, "rationale": "Proposes a governance framework for research on deceptive AI and its risk levels. This is AI safety and governance.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399042091, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Ethical analysis of health recommender systems based on the autonomy principle. It is AI ethics and governance.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214017973, "categories": [], "primary": null, "confidence": 0.55, "rationale": "Case study of faculty using generative AI in academic program design. It is education and ethics practice, not core computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 2803572453, "categories": ["information_retrieval"], "primary": "information_retrieval", "confidence": 0.35, "rationale": "Empirical study of verbatim content copying networks among news producers; web mining flavored, borderline CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3118441209, "categories": ["privacy_preserving_computation"], "primary": "privacy_preserving_computation", "confidence": 0.5, "rationale": "Ethics and privacy of web analytics personal data collection, with policy measures.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396786581, "categories": ["ai_safety", "ml_theory"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Shapley value-based feature detection for fairness and disparity in ML models.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161719656, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.6, "rationale": "Position paper on AI agents generating hypotheses, papers and reviews and the epistemology of machine-generated science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164757242, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.35, "rationale": "Conceptual risk governance framework for algorithmic recommendation in education; governance focus, weakly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 7119498334, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Comparative analysis of AI governance and regulation policy texts across US, EU, China.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221224735, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Community service workshop on using NotebookLM in religious education; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400082893, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Survey study of students' attitudes toward AI in education; educational research, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4394613885, "categories": [], "primary": null, "confidence": 0.65, "rationale": "Educational policy study on AI professional community initiative in Oman; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415318303, "categories": ["ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.92, "rationale": "Benchmark evaluating LLM harms such as delusion reinforcement and harm enablement, a safety evaluation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386791501, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.55, "rationale": "Proposes AI supplemental questions for IRB review to mitigate AI risks; governance of AI research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411990784, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Questionnaire survey of student attitudes toward AI; weak fit to computer science, closest to human-AI perception studies.", "model": "claude-sonnet-5-5"}
++{"work_id": 4220714491, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Curriculum analysis of informatics and society courses; computing education, not covered by the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3096804093, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.3, "rationale": "Survey and recommendations on AI implementation in African governments; policy/governance orientation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405384079, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "Interview study of researchers' trauma and an ethics of care; social computing flavored, marginal CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3093512442, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.35, "rationale": "Editorial guidance on bias, generalizability and reporting quality for healthcare AI manuscripts.", "model": "claude-sonnet-5-5"}
++{"work_id": 7147573305, "categories": ["llm", "privacy_preserving_computation"], "primary": "llm", "confidence": 0.75, "rationale": "Audits membership inference as a diagnostic for memorization contamination in LLM financial forecasting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415829954, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Survey of AI tool usage and competence among Nigerian academics; education/policy study, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392109933, "categories": ["autonomous_scientific_discovery"], "primary": "autonomous_scientific_discovery", "confidence": 0.4, "rationale": "Discussion of epistemology and agency in AI-driven research processes.", "model": "claude-sonnet-5-5"}
++{"work_id": 7211974235, "categories": [], "primary": null, "confidence": 0.5, "rationale": "General essay on AI and human communication in education and society; not core CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3110765181, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.6, "rationale": "Position paper on societal harms of AI (bias, privacy, surveillance) and the need for interdisciplinary work.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399365038, "categories": ["llm", "ai_safety"], "primary": "ai_safety", "confidence": 0.65, "rationale": "Algorithm audit of ChatGPT content moderation guardrails and their biases.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386555180, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Discusses limits of XAI and responsible AI including fairness, privacy and contestability.", "model": "claude-sonnet-5-5"}
++{"work_id": 4283010037, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.4, "rationale": "Overview of how network algorithms such as ranking and recommendation affect minorities, framed as fairness and inequality.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412611638, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Conceptual and philosophical paper on analogical reasoning as a core AGI capability; no clear category fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 4290803082, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Legal and policy analysis of data governance and Indigenous data sovereignty; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7204733568, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Education policy survey of stakeholder preferences; not core computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162150262, "categories": ["ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.75, "rationale": "Behavioral safety evaluation of open-source LLMs under authority pressure in a Milgram-style experiment.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390091701, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Argues for moral dimensions beyond fairness in ethical AI assessment and design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399331704, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Fuzzy decision-making and balanced scorecard study of AI adoption in finance; business research, not core computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134710523, "categories": ["ai_safety", "hci"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Conceptual mental model of explanation for AI decisions (XAI, user trust); interpretability plus human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4399198750, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.8, "rationale": "Fairness measures and bias mitigation for tabular datasets with multiple protected attributes, tied to the AI Act.", "model": "claude-sonnet-5-5"}
++{"work_id": 7135082306, "categories": ["ai_safety", "llm"], "primary": "ai_safety", "confidence": 0.88, "rationale": "Mechanistic interpretability of safety neurons in Llama-3-70B.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405576526, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.7, "rationale": "Survey of AI regulation and governance in New Zealand.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411795627, "categories": ["ai_safety", "hci"], "primary": "ai_safety", "confidence": 0.5, "rationale": "Critical reflection on AI bias and algorithmic justice in inclusive math education, with accessibility concerns.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164020700, "categories": ["generative_ai"], "primary": "generative_ai", "confidence": 0.35, "rationale": "Discusses ethics, copyright and originality in AI-generated art and design; weak CS content.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134826009, "categories": ["ai_safety"], "primary": "ai_safety", "confidence": 0.4, "rationale": "Qualitative AI governance framework for Islamic higher education institutions.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165690142, "categories": ["ai_safety", "hci"], "primary": "ai_safety", "confidence": 0.85, "rationale": "Red teaming of ChatGPT with faith leaders; LLM safety and accountability, with a participatory workshop study.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117162136, "categories": ["ai_safety", "llm", "ai_agents"], "primary": "ai_safety", "confidence": 0.9, "rationale": "Evaluation of chain-of-thought monitorability of reasoning models for safe oversight of agents.", "model": "claude-sonnet-5-5"}
++{"work_id": 2997003061, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.5, "rationale": "Model-based safety design (STPA) integrated into the aircraft systems development process.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133303346, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Analog/digital circuit design for piezoelectric energy harvesting interface (ISSCC); closest fit is low-power VLSI design.", "model": "claude-sonnet-5-5"}
++{"work_id": 2950925562, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Pure geometric measure theory mathematics, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3115601909, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Auction mechanism design for P2P energy trading with truthfulness and efficiency analysis, i.e. algorithmic game theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4384570508, "categories": ["distributed_computing", "computer_networks"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Scalable archiving system framework with load balancing, fault tolerance, and data placement; also hybrid switching.", "model": "claude-sonnet-5-5"}
++{"work_id": 3187190559, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Multi-objective workflow scheduling of cloud resources using an evolutionary algorithm with a neural network.", "model": "claude-sonnet-5-5"}
++{"work_id": 7128410034, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.6, "rationale": "Algorithms for computing the rational univariate representation of parametric polynomial systems, with degree and height bounds (computer algebra).", "model": "claude-sonnet-5-5"}
++{"work_id": 3006062179, "categories": ["distributed_computing", "ai_accelerators"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Performance simulation of ML systems via profiling, considering parallelization strategies in distributed training.", "model": "claude-sonnet-5-5"}
++{"work_id": 2860211826, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Fault-tolerant fractional-order control of an isotope separation process; control engineering, not core computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133367970, "categories": ["distributed_computing", "deep_learning_architectures"], "primary": "distributed_computing", "confidence": 0.7, "rationale": "Deep learning on logs for failure prediction and fault tolerance in HPC systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 2907696829, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Microgrid energy management with model predictive control; power/control engineering, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 2918759589, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Special issue editorial on HPC and distributed systems covering GPU computing, parallel applications and cloud computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2916417520, "categories": ["distributed_computing", "compilers"], "primary": "distributed_computing", "confidence": 0.55, "rationale": "Gecko is a directive-based programming model for multi-GPU and heterogeneous HPC systems, with a source-to-source transformation and a runtime library.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413272460, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.93, "rationale": "Blockchain framework addressing the trilemma through proof-of-capacity consensus, sharding and decentralized storage.", "model": "claude-sonnet-5-5"}
++{"work_id": 2805585446, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Engineering education study, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409660083, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Power electronics inverter control, which is electrical engineering rather than computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7161207400, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Seismology instrument announcement with no computer science content.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400690093, "categories": ["blockchain"], "primary": "blockchain", "confidence": 0.92, "rationale": "Compares optimistic and zero-knowledge layer-2 rollups as blockchain scalability solutions.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388580994, "categories": ["computer_architecture", "distributed_computing"], "primary": "computer_architecture", "confidence": 0.65, "rationale": "Performance evaluation of the high-bandwidth-memory Xeon MAX CPU on bandwidth-bound HPC applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 2904849773, "categories": ["software_engineering"], "primary": "software_engineering", "confidence": 0.4, "rationale": "Builds a business training platform on workflow management systems, covering framework design and implementation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4401163816, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Signal processing for proton magnetometer instrumentation, not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7201761843, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.75, "rationale": "Geo-distributed Kafka clusters with replication and failover is a distributed systems topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 3152708525, "categories": ["distributed_computing", "computer_networks"], "primary": "distributed_computing", "confidence": 0.55, "rationale": "Optimistic synchronization and server selection for delay in distributed processing of communication services.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200525533, "categories": ["nlp"], "primary": "nlp", "confidence": 0.6, "rationale": "Deep learning text analysis of social media messages for suicide risk, a text classification task.", "model": "claude-sonnet-5-5"}
++{"work_id": 3038439485, "categories": [], "primary": null, "confidence": 0.97, "rationale": "Educational pop-up book development for reading skills is not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 2972427306, "categories": ["distributed_computing", "ai_accelerators"], "primary": "distributed_computing", "confidence": 0.7, "rationale": "GPU acceleration of Map-Reduce/Spark for iterative ML, with middleware and scheduling insights.", "model": "claude-sonnet-5-5"}
++{"work_id": 3014911997, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control-theoretic synchronization of fractional-order memristive networks is mainly applied math and control, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2998437218, "categories": ["cloud_computing"], "primary": "cloud_computing", "confidence": 0.8, "rationale": "Ontology-based cloud monitoring system and resource management with fault self-diagnosis.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416926542, "categories": ["cloud_computing", "sustainable_computing", "algorithms"], "primary": "cloud_computing", "confidence": 0.75, "rationale": "Energy-aware multi-objective job scheduling in green cloud datacenters using a hybrid metaheuristic.", "model": "claude-sonnet-5-5"}
++{"work_id": 2903036202, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control design for synchronization of Lur'e systems via LMIs is control theory, outside CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4286377584, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.35, "rationale": "Empirical Bayesian estimation for InSAR coherence is a statistical signal-processing method for remote sensing, a weak fit to the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 7131086620, "categories": [], "primary": null, "confidence": 0.6, "rationale": "A survey of radiologists' opinions on AI in breast imaging; it is a clinical opinion survey, not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3045355300, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Control-theory work on adaptive consensus for linear multiagent systems; it falls outside the CS ontology categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405718414, "categories": ["software_engineering", "distributed_computing"], "primary": "software_engineering", "confidence": 0.7, "rationale": "Retrospective on dynamic change management and safe updating of running distributed systems, published in TSE.", "model": "claude-sonnet-5-5"}
++{"work_id": 2903731864, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Statistical simulation study of count data models; it is statistics rather than CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4200552233, "categories": ["distributed_computing", "software_engineering"], "primary": "distributed_computing", "confidence": 0.6, "rationale": "Selective instrumentation to reduce overhead in Python performance monitoring with Score-P, which is HPC performance analysis tooling.", "model": "claude-sonnet-5-5"}
++{"work_id": 7216249722, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Indonesian language education research on an e-module for short-story writing; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2891039752, "categories": ["cloud_computing", "distributed_computing", "operating_systems"], "primary": "cloud_computing", "confidence": 0.65, "rationale": "Hypervisor-level checkpoint-restart for fault tolerance in virtual machine clusters running MPI in the cloud.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403092717, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.4, "rationale": "Parallel processing of geographic data on mobile devices, which is on-device computation.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407784105, "categories": ["distributed_computing", "deep_learning_architectures"], "primary": "distributed_computing", "confidence": 0.7, "rationale": "Parallel GNN training on large graphs across heterogeneous platforms, using graph partitioning and workload balancing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411552467, "categories": ["blockchain", "distributed_computing"], "primary": "blockchain", "confidence": 0.8, "rationale": "Improved PBFT consensus algorithm for medical device traceability.", "model": "claude-sonnet-5-5"}
++{"work_id": 2887272015, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.88, "rationale": "HPC resource management and job malleability via an interface integrated with OpenMP and MPI.", "model": "claude-sonnet-5-5"}
++{"work_id": 3123252765, "categories": ["robotics"], "primary": "robotics", "confidence": 0.4, "rationale": "Vibration suppression in servo motion control using input shaping and adaptive control; closest fit is robotics/motion control, though not a clean match.", "model": "claude-sonnet-5-5"}
++{"work_id": 3093694272, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Information systems management study of IT affordance and organisational innovation; not CS research in the ontology's sense.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411584530, "categories": ["reinforcement_learning", "cloud_computing"], "primary": "reinforcement_learning", "confidence": 0.75, "rationale": "Distributional RL method for Pod (Kubernetes) scheduling in satellite clusters; spans RL and cloud-style resource scheduling.", "model": "claude-sonnet-5-5"}
++{"work_id": 4383532084, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Erasure-coded distributed archival storage system on EOS for the WLCG, focused on data protection and storage design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3048509415, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Literature review on manufacturing quality management (Lean, Six Sigma, ZDM); not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414878360, "categories": ["programming_languages", "distributed_computing"], "primary": "programming_languages", "confidence": 0.8, "rationale": "A DSL for mapping distributed programs, built on the Legion runtime.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410853603, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "AI-assisted system for HPC remote access and resource scheduling, with a RoBERTa chatbot as a supporting component.", "model": "claude-sonnet-5-5"}
++{"work_id": 4407673169, "categories": ["distributed_computing", "cloud_computing"], "primary": "distributed_computing", "confidence": 0.8, "rationale": "Self-healing framework for fault tolerance and recovery in distributed cloud systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011113897, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.85, "rationale": "Multi-stage distributed big-data processing framework evaluated across connection topologies.", "model": "claude-sonnet-5-5"}
++{"work_id": 3128655982, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.75, "rationale": "Improved DDPG for obstacle avoidance path; RL algorithm improvement with a navigation application.", "model": "claude-sonnet-5-5"}
++{"work_id": 2996458828, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "Meta-RL with subtask graph inference and exploration reward.", "model": "claude-sonnet-5-5"}
++{"work_id": 4403461025, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.5, "rationale": "Adaptive dynamic programming (an RL method) for wind turbine MPPT control; control application, only loosely RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160928702, "categories": ["ai_agents", "reinforcement_learning", "llm"], "primary": "ai_agents", "confidence": 0.8, "rationale": "LLM agents with external skill bank managed during agentic RL training.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162259189, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Evaluates task complexity metrics in RL on robotic manipulation setups.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388561956, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "Hierarchical RL (MAXQ-P extension) compared with Q-learning on a game.", "model": "claude-sonnet-5-5"}
++{"work_id": 4285172767, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.88, "rationale": "Frame-skipping in deep RL; uses hidden observations to improve trained agent performance.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400838813, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.82, "rationale": "DDPG-based deep RL for portfolio management with reward design and GCN/TCN features.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416874462, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.93, "rationale": "Off-policy average-reward actor-critic algorithm.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138837252, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Faculty highlight on reward-free exploration, imitation and RL without explicit rewards.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412985292, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "PPO variant with reward shaping for robot path planning in grid world.", "model": "claude-sonnet-5-5"}
++{"work_id": 4384135373, "categories": ["reinforcement_learning", "ai_safety"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Explainable RL using fuzzy rules and generative visualization; RL focus with interpretability.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413402209, "categories": ["robotics", "reinforcement_learning"], "primary": "robotics", "confidence": 0.65, "rationale": "UAV path planning combining Q-learning/MDP with improved artificial potential field.", "model": "claude-sonnet-5-5"}
++{"work_id": 4411020042, "categories": ["reinforcement_learning", "ml_theory"], "primary": "ml_theory", "confidence": 0.7, "rationale": "Combinatorial logistic bandits with regret bounds; online learning theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414230988, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.65, "rationale": "SAC with PER for autonomous vehicle collision avoidance in CARLA.", "model": "claude-sonnet-5-5"}
++{"work_id": 3177555056, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "Dissertation on general value functions, TD learning, successor representations.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400812532, "categories": ["robotics", "reinforcement_learning"], "primary": "robotics", "confidence": 0.8, "rationale": "Mobile robot navigation and control with ROS, SLAM and actor-critic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405571822, "categories": ["reinforcement_learning", "robotics"], "primary": "robotics", "confidence": 0.6, "rationale": "Comparative DRL algorithms for tilt-rotor UAV transition control.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214803104, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.93, "rationale": "Thesis on multi-objective RL methods for continuous control.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392858396, "categories": ["hci"], "primary": "hci", "confidence": 0.3, "rationale": "AI education game-based learning activities; weak fit, nearest is user-centered interaction/education study.", "model": "claude-sonnet-5-5"}
++{"work_id": 3040890460, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Contextual bandit variant with missing rewards; bandits fall under RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405358450, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Musicology open educational resources review; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388579628, "categories": ["reinforcement_learning", "robotics"], "primary": "robotics", "confidence": 0.7, "rationale": "SAC-based parking path planning and tracking for autonomous vehicles; robotics application with DRL.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160554706, "categories": ["reinforcement_learning", "computational_biology"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Review of dopamine reward prediction error hypothesis linking neuroscience and TD learning; computational neuroscience.", "model": "claude-sonnet-5-5"}
++{"work_id": 3216293753, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Active inference planning via tree search; decision-making and planning theory akin to RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313459177, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Evolutionary deep RL with SAC for Volt-VAR control in power distribution networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 3014616672, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Monte Carlo Q-value function for Dec-POMDP multi-agent decision making and heuristic search.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213255109, "categories": ["reinforcement_learning", "algorithms"], "primary": "reinforcement_learning", "confidence": 0.55, "rationale": "MDP abstraction algorithm (K-MDP) for interpretable policies; decision process methods.", "model": "claude-sonnet-5-5"}
++{"work_id": 7151096044, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "MARL stabilization method SD-MAPPO using self-distillation regularization.", "model": "claude-sonnet-5-5"}
++{"work_id": 7167869540, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.45, "rationale": "Dynamic programming for Dec-POMDPs with delayed sharing; decentralized control/decision theory, abstract unavailable.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408611127, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.9, "rationale": "Studies how state-space representation affects DQN agents in Snake; core RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 3208966735, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.95, "rationale": "Intrinsic-reward exploration method for cooperative multi-agent RL with sparse rewards.", "model": "claude-sonnet-5-5"}
++{"work_id": 4385764347, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "State abstraction framework for RL with structural information principles.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412795647, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Multi-agent deep RL (QMIX variant) for truck platoon coordination; the RL method dominates, with some link to autonomous vehicles.", "model": "claude-sonnet-5-5"}
++{"work_id": 7123339254, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Theoretical analysis of strong duality for constrained average-reward MDPs with intermittent observations; RL/MDP theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221661943, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.7, "rationale": "Actor-critic RL for distributed optimal output regulation of multi-agent systems; control-flavored but RL-based.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391768605, "categories": ["llm", "reinforcement_learning", "hci"], "primary": "llm", "confidence": 0.6, "rationale": "User study of ChatGPT helping novices design RL state and reward functions for traffic control; covers LLM use, RL formulation and human-AI interaction.", "model": "claude-sonnet-5-5"}
++{"work_id": 3170511602, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.95, "rationale": "Cooperative exploration method for multi-agent deep RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414610855, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.6, "rationale": "Q-learning applied to personalised exercise recommendation for vocabulary teaching; applied RL.", "model": "claude-sonnet-5-5"}
++{"work_id": 4406073402, "categories": ["reinforcement_learning", "ai_safety"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Label-efficient concept bottleneck policies for interpretable RL; the training algorithm is the main contribution, with an interpretability angle.", "model": "claude-sonnet-5-5"}
++{"work_id": 3174702694, "categories": ["computer_vision", "reinforcement_learning", "ai_agents", "robotics"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Embodied agent actively selects views to improve semantic segmentation, trained with deep RL in a simulator.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402437985, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Two-tier multi-agent RL for strategy and tactics in competitive games.", "model": "claude-sonnet-5-5"}
++{"work_id": 2793827368, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.55, "rationale": "Adaptive dynamic programming with critic networks for optimal control, which is approximate RL in the control setting.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400526476, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Offline-to-online RL for quadrotor aerial docking.", "model": "claude-sonnet-5-5"}
++{"work_id": 7213353131, "categories": ["ai_agents", "autonomous_scientific_discovery"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Recursive self-improving exploration for autonomous agents that discover solutions, using a replay simulator to refine exploration policies.", "model": "claude-sonnet-5-5"}
++{"work_id": 3217267205, "categories": ["reinforcement_learning", "ml_theory"], "primary": "ml_theory", "confidence": 0.4, "rationale": "Probabilistic hierarchical Bayesian simulator of spatial demand, used to train RL policies with Deep Q-learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7127355566, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.92, "rationale": "New target update method for DQN with convergence analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 7117961497, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Actor-critic RL with Lyapunov certificates for stabilizing control.", "model": "claude-sonnet-5-5"}
++{"work_id": 4393861500, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Stochastic control and PDE free-boundary analysis in mathematical finance and economics, which is not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 6884998335, "categories": ["reinforcement_learning", "ml_theory"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Convergence and regret bounds for q-learning in continuous-time controlled diffusions.", "model": "claude-sonnet-5-5"}
++{"work_id": 4220992898, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Neuromorphic spiking sensory neuron hardware (memristor) for in-sensor computing; hardware for neural computation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3011341896, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.85, "rationale": "Sparsity of compressed CNNs analyzed on a hardware accelerator model.", "model": "claude-sonnet-5-5"}
++{"work_id": 3012645150, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.85, "rationale": "Photonic CNN architecture for integrated optical deep learning processors.", "model": "claude-sonnet-5-5"}
++{"work_id": 4388847637, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Review of optical neural networks as hardware, with 5G/optical communication context.", "model": "claude-sonnet-5-5"}
++{"work_id": 3047966047, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "Defect characterization of spintronic neuromorphic circuits; hardware testing for AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416096548, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Curriculum/teaching reform paper, not core CS research contribution in the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414291128, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.92, "rationale": "Fabricated on-chip photonic neural network for inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400641184, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "Spin-torque oscillator reservoir computing hardware for energy-efficient neuromorphic computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7133525837, "categories": ["ml_theory"], "primary": "ml_theory", "confidence": 0.5, "rationale": "Mathematical analysis of synchronization dynamics in memristive Hopfield networks; theoretical.", "model": "claude-sonnet-5-5"}
++{"work_id": 2989832056, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.75, "rationale": "Photonic laser-based reservoir computing hardware and its computational capacity.", "model": "claude-sonnet-5-5"}
++{"work_id": 7138209559, "categories": ["computer_architecture", "ai_accelerators"], "primary": "computer_architecture", "confidence": 0.5, "rationale": "Memristive neuromorphic circuit design with firing dynamics and PCB validation; hardware-level neuromorphic building block.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312821358, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Analysis of in-memory vector-matrix multiplication on ReRAM crossbars, a core kernel for neuromorphic and AI hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387594915, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.45, "rationale": "Photonic neural networks are AI hardware, though the abstract is empty and the title is the only evidence.", "model": "claude-sonnet-5-5"}
++{"work_id": 7165771277, "categories": ["deep_learning_architectures", "ai_accelerators"], "primary": "deep_learning_architectures", "confidence": 0.7, "rationale": "Spiking transformer architecture for modulation recognition, evaluated on a neuromorphic chip for energy savings.", "model": "claude-sonnet-5-5"}
++{"work_id": 4392976059, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Brief review of an open science training resource; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3193775649, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.35, "rationale": "Memristive synaptic device interface for neural signals and neuroprosthetics; hardware-oriented neuromorphic work, a weak fit.", "model": "claude-sonnet-5-5"}
++{"work_id": 2912636021, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.85, "rationale": "Low-power QCA MAC unit circuit design with power analysis; digital hardware and VLSI design.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408280919, "categories": ["cybersecurity", "applied_cryptography"], "primary": "applied_cryptography", "confidence": 0.4, "rationale": "Memristive coupled chaotic neural networks, with a chaos-based DNA image encryption algorithm as the application.", "model": "claude-sonnet-5-5"}
++{"work_id": 3039564587, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Device physics characterization of h-BN memristors under a magnetic field, with no computational framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2955631546, "categories": [], "primary": null, "confidence": 0.55, "rationale": "Materials and device physics model of negative differential resistance, mostly outside computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 2846788133, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.4, "rationale": "Analog chaotic oscillator circuit design with memristor emulation; closest fit is digital/analog hardware design, a weak match.", "model": "claude-sonnet-5-5"}
++{"work_id": 3191516184, "categories": ["ai_accelerators", "computer_architecture"], "primary": "computer_architecture", "confidence": 0.45, "rationale": "Tutorial on programmable photonic processors and their technology stack; general-purpose photonic computing hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 2902787482, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.6, "rationale": "CMOS-memristive analog current mirror circuit design for low power and area, i.e. VLSI circuit design.", "model": "claude-sonnet-5-5"}
++{"work_id": 3091740059, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.65, "rationale": "Memristive crossbar threshold logic circuit design with power and variability analysis; reconfigurable hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 7084748818, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.9, "rationale": "Neuromorphic self-attention hardware core exploiting sparsity for energy efficiency.", "model": "claude-sonnet-5-5"}
++{"work_id": 4405521075, "categories": ["computer_architecture", "computational_biology"], "primary": "computer_architecture", "confidence": 0.55, "rationale": "FPGA digital hardware realization of a synaptic calcium dynamics model, with a neuroscience modelling aspect.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206136645, "categories": ["computer_architecture"], "primary": "computer_architecture", "confidence": 0.8, "rationale": "VHDL-based design space exploration and EDA framework for memristive crossbar arrays.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225485808, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.8, "rationale": "All-optical nonlinear activation device for optical neural networks, i.e. photonic AI hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410384882, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.88, "rationale": "Integrated all-optical diffractive neural network chip for low-latency inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391086735, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.55, "rationale": "Silicon photonic coherent crossbar for universal linear optics, i.e. photonic matrix computation hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4323823183, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.5, "rationale": "Simulation and training of optical neural network systems; hardware-aware modelling for photonic AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 4391451283, "categories": ["ai_accelerators", "edge_computing"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Diffractive photonic computing units for energy-efficient EEG seizure classification as edge systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3047123852, "categories": ["computer_architecture", "ai_accelerators"], "primary": "computer_architecture", "confidence": 0.7, "rationale": "Behavioral VHDL modelling of memristive crossbar for in-memory computing; EDA and simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 3143061595, "categories": ["ai_accelerators", "edge_computing"], "primary": "ai_accelerators", "confidence": 0.75, "rationale": "Overview of AI chip technologies and design flow for AGI chips, including edge AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 3153206747, "categories": ["ai_accelerators", "computer_networks"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Microcomb-based optical convolutional accelerator for neural networks plus optical data transmission.", "model": "claude-sonnet-5-5"}
++{"work_id": 4408740885, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Photonic hardware for probabilistic Bayesian neural network inference.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413066757, "categories": ["ai_accelerators", "computer_architecture"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Optical network-on-chip interconnect and mapping algorithms for memristor arrays in neuromorphic chips.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134888657, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.75, "rationale": "Programmable optical neural network (optoelectronic reservoir computing) hardware for efficient AI.", "model": "claude-sonnet-5-5"}
++{"work_id": 4240155423, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.85, "rationale": "Photonic convolutional accelerator for optical neural networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 4413807734, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Optical synapse device for neuromorphic computing hardware.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412921988, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.7, "rationale": "Review of photonic reservoir computing, brain-inspired hardware for efficient AI computation.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164587856, "categories": ["ai_accelerators", "deep_learning_architectures"], "primary": "ai_accelerators", "confidence": 0.85, "rationale": "Photonic spiking GNN with silicon photonic hardware experiment and energy efficiency metrics.", "model": "claude-sonnet-5-5"}
++{"work_id": 6929467513, "categories": ["deep_learning_architectures", "ml_theory"], "primary": "deep_learning_architectures", "confidence": 0.5, "rationale": "Generalized threshold logic neuron model (receptron) with formal analysis of its selectivity.", "model": "claude-sonnet-5-5"}
++{"work_id": 4400662090, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.6, "rationale": "Photonic iterative processor for matrix inversion addressing I/O bottleneck, with reservoir training and MIMO applications.", "model": "claude-sonnet-5-5"}
++{"work_id": 4366823742, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.88, "rationale": "Photonic tensor core with MZI chip for matrix multiplication in photonic neural networks.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171493874, "categories": ["ai_accelerators"], "primary": "ai_accelerators", "confidence": 0.3, "rationale": "Photonic chip with self-configuring networks learning coherency matrices; borderline computational framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 4386996793, "categories": [], "primary": null, "confidence": 0.9, "rationale": "InSAR geophysics study of fault deformation, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2801359411, "categories": ["computational_biology"], "primary": "computational_biology", "confidence": 0.7, "rationale": "Computational neuroscience modeling of synaptic plasticity with hidden Markov models.", "model": "claude-sonnet-5-5"}
++{"work_id": 2946793717, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Circuit-level memristor device dynamics analysis, essentially electrical engineering rather than computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 3112050902, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Control-theoretic synchronization of memristive neural networks, mainly control engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 4383046458, "categories": ["robotics", "ml_theory"], "primary": "robotics", "confidence": 0.4, "rationale": "Data-driven MPC with safety guarantees for nonlinear systems (control); closest fits are control/robotics and learning-based methods, but the fit is loose.", "model": "claude-sonnet-5-5"}
++{"work_id": 2804990859, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.92, "rationale": "Proves #P-completeness of counting homomorphisms for 3-manifold invariants.", "model": "claude-sonnet-5-5"}
++{"work_id": 2795888755, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Pure algebra (monoid acts); not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7205015363, "categories": ["computational_complexity", "computational_biology"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Automata theory and formal languages applied to DNA splicing systems, which falls under automata theory and DNA computing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2979317633, "categories": [], "primary": null, "confidence": 0.6, "rationale": "NIR spectroscopy for skin AGE estimation is biomedical measurement, not core CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3037346090, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.45, "rationale": "Numerical approximation (inverse-free Kalman filter using block diagonally dominant matrices) to reduce computational cost.", "model": "claude-sonnet-5-5"}
++{"work_id": 2947814064, "categories": ["scientific_computing", "computational_biology"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Simulation of biological processes using movable cellular automata.", "model": "claude-sonnet-5-5"}
++{"work_id": 4312278247, "categories": [], "primary": null, "confidence": 0.5, "rationale": "Signal processing measure for time-varying filter banks; no CS category fits well.", "model": "claude-sonnet-5-5"}
++{"work_id": 2928354599, "categories": ["computational_biology", "scientific_computing"], "primary": "computational_biology", "confidence": 0.55, "rationale": "Cellular automata model of morphogenesis, a computational developmental biology simulation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2791221456, "categories": [], "primary": null, "confidence": 0.4, "rationale": "Applied fuzzy Tsukamoto fire-risk prediction; weak CS contribution, no category fits clearly.", "model": "claude-sonnet-5-5"}
++{"work_id": 4410227679, "categories": ["deep_learning_architectures", "robotics"], "primary": "robotics", "confidence": 0.4, "rationale": "Deep-learning-enhanced IMM filter for maneuvering target tracking; a signal processing/tracking application with weak fit to the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 2936835067, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.9, "rationale": "Unique Games hardness of approximating cumulative pebbling cost; hardness of approximation result.", "model": "claude-sonnet-5-5"}
++{"work_id": 4323338639, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "ADMM-based optimization algorithm for regularized minimax approximation, applied to filter design and neural network training.", "model": "claude-sonnet-5-5"}
++{"work_id": 3090906110, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.4, "rationale": "Reaction systems transition graph structure; a theoretical natural computing/automata-style result.", "model": "claude-sonnet-5-5"}
++{"work_id": 3174631184, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.85, "rationale": "Fixed point construction for tilings and cellular automata using Turing machine simulation; computability and automata theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 7134076622, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.35, "rationale": "Operations research survey of optimization models (MIP) for offshore wind farms; combinatorial optimization, but the domain is outside computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387356133, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.95, "rationale": "PSPACE-completeness proof for the game Turning Tiles.", "model": "claude-sonnet-5-5"}
++{"work_id": 2567311606, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.7, "rationale": "Algorithmic randomness and infinite time Turing machines; computability theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 7125235431, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Geoscience study using PLS regression on hyperspectral and petrophysical data; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2787143214, "categories": ["formal_verification", "computational_complexity"], "primary": "formal_verification", "confidence": 0.6, "rationale": "Decidability of the weak monadic theory of successor; logic and automata theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162015013, "categories": [], "primary": null, "confidence": 0.95, "rationale": "Management/education survey study using SEM-PLS; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4361029879, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.4, "rationale": "Distributed MPC with local information exchange and guarantees; control theory, nearest fit is distributed algorithms.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387624329, "categories": ["deep_learning_architectures"], "primary": "deep_learning_architectures", "confidence": 0.6, "rationale": "Proposes an efficient Informer-based transformer architecture with sparse attention and distilling for converter modeling.", "model": "claude-sonnet-5-5"}
++{"work_id": 2794110548, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.3, "rationale": "Granular computing and cellular automata soft computing model; weak fit to computational models/algorithms.", "model": "claude-sonnet-5-5"}
++{"work_id": 2990666756, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Classical control engineering (IMC design for MIMO systems); no CS category fits.", "model": "claude-sonnet-5-5"}
++{"work_id": 2888657220, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.4, "rationale": "Adaptive filtering algorithm (proportionate NLMS) for sparse system identification; signal processing algorithm design.", "model": "claude-sonnet-5-5"}
++{"work_id": 7221079107, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.95, "rationale": "Shows barriers implying NP \u2286 P/poly for affine isolation; pure complexity theory.", "model": "claude-sonnet-5-5"}
++{"work_id": 2968152157, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Geophysical exploration application; AI used incidentally in geoscience, not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4414170577, "categories": ["robotics"], "primary": "robotics", "confidence": 0.5, "rationale": "Safety-critical control with CBFs for an underactuated hovercraft, trajectory tracking and obstacle avoidance.", "model": "claude-sonnet-5-5"}
++{"work_id": 2905468286, "categories": [], "primary": null, "confidence": 0.98, "rationale": "Microbiology laboratory equipment; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 3110879399, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.95, "rationale": "Algebraic circuit complexity, VNP and permanent hardness.", "model": "claude-sonnet-5-5"}
++{"work_id": 4396536343, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Radar signal processing parameter estimation; not core CS in this ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 4288109977, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.5, "rationale": "Cellular automata simulation of corrosion; abstract minimal, a computational modelling topic.", "model": "claude-sonnet-5-5"}
++{"work_id": 4417095717, "categories": ["computational_complexity", "deep_learning_architectures"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Uses infinite time Turing machines to analyze transformers and proposes a new computational paradigm.", "model": "claude-sonnet-5-5"}
++{"work_id": 2955693445, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.5, "rationale": "Root separation bounds for polynomials, computer algebra flavored.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409570626, "categories": ["quantum_computing"], "primary": "quantum_computing", "confidence": 0.3, "rationale": "Probabilistic cellular automata as quantum systems; mostly physics with information framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 2888883998, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.9, "rationale": "New automata model, descriptional complexity, language hierarchies.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214073407, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Finite basis problem for semirings is algebra, not clearly CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2809671584, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Number theory experimental mathematics, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 4206034737, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Application of standard ML to remote sensing climate data; main contribution is geoscience.", "model": "claude-sonnet-5-5"}
++{"work_id": 6967003509, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.97, "rationale": "Characterizes lossy catalytic space classes in terms of ordinary catalytic space, a complexity class result.", "model": "claude-sonnet-5-5"}
++{"work_id": 2738145410, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.7, "rationale": "Stochastic mixed integer programming with tight formulations and cuts is combinatorial optimisation.", "model": "claude-sonnet-5-5"}
++{"work_id": 2952324823, "categories": ["scientific_computing"], "primary": "scientific_computing", "confidence": 0.6, "rationale": "Numerical simulation and stability analysis of a reactor model, computational chemical engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 2962975699, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.9, "rationale": "Classifies the complexity of infinite-domain VCSPs, with tractability and NP-hardness results.", "model": "claude-sonnet-5-5"}
++{"work_id": 7169788258, "categories": ["computer_vision", "deep_learning_architectures"], "primary": "computer_vision", "confidence": 0.7, "rationale": "YOLO-based detection and recognition on time-frequency images of radar signals, with architectural modifications.", "model": "claude-sonnet-5-5"}
++{"work_id": 2995807397, "categories": ["computational_complexity"], "primary": "computational_complexity", "confidence": 0.93, "rationale": "Proves strong NP-hardness of a clustering problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 4409702353, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Survey study of MOOC adoption in education using PLS-SEM, not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7148239571, "categories": ["applied_cryptography", "computational_complexity"], "primary": "computational_complexity", "confidence": 0.5, "rationale": "Thesis on cryptography and Kolmogorov complexity, likely about the relationship between cryptographic hardness and complexity; the abstract is minimal.", "model": "claude-sonnet-5-5"}
++{"work_id": 3086506771, "categories": [], "primary": null, "confidence": 0.75, "rationale": "Pure algebra of ordered semihypergroups, with no computational framing.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154114167, "categories": ["edge_computing"], "primary": "edge_computing", "confidence": 0.3, "rationale": "IoT sensor-based fuzzy Mamdani classification of water quality, a weak fit; it is applied fuzzy logic with IoT data.", "model": "claude-sonnet-5-5"}
++{"work_id": 4382237584, "categories": ["robotics", "ai_agents"], "primary": "robotics", "confidence": 0.6, "rationale": "Decision-making for automated vehicle interacting with pedestrians; autonomous vehicle system with intention prediction.", "model": "claude-sonnet-5-5"}
++{"work_id": 4313639087, "categories": ["distributed_computing"], "primary": "distributed_computing", "confidence": 0.3, "rationale": "Control of networked multi-agent systems with consensus and delay compensation; closest fit is distributed coordination, weak.", "model": "claude-sonnet-5-5"}
++{"work_id": 2978699180, "categories": ["computational_biology", "formal_verification", "nlp"], "primary": "computational_biology", "confidence": 0.5, "rationale": "Mines literature via NLP to build and simulate biological network models, checked with statistical model checking.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416034050, "categories": ["ai_agents", "ai_safety", "cybersecurity"], "primary": "ai_agents", "confidence": 0.7, "rationale": "Security framework with attacks and defenses for LLM-based multi-agent systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4387171855, "categories": ["ai_agents", "algorithms"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Numeric AI planning heuristic improving LM-cut for optimal planning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7202147256, "categories": ["reinforcement_learning", "ai_agents", "ai_safety"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Preference-based reward modeling for multi-agent RL to shape emergent behavior and align with social objectives.", "model": "claude-sonnet-5-5"}
++{"work_id": 7171217691, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Tourism labour and social sustainability commentary, not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4281774473, "categories": [], "primary": null, "confidence": 0.3, "rationale": "Thesis on feature extraction and selection for efficient time series classification; abstract text mostly acknowledgements, but it is a mining/ML method; no clear ontology category.", "model": "claude-sonnet-5-5"}
++{"work_id": 7212269269, "categories": ["reinforcement_learning", "ai_agents"], "primary": "reinforcement_learning", "confidence": 0.5, "rationale": "Update admission auditing for continual learning of embodied agents with statistical gating of policy updates.", "model": "claude-sonnet-5-5"}
++{"work_id": 4225668571, "categories": ["nlp", "hci"], "primary": "nlp", "confidence": 0.5, "rationale": "Overview of chatbots and conversational agents for education and coaching.", "model": "claude-sonnet-5-5"}
++{"work_id": 3095072217, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Multi-agent simulation platform for mobility scenarios; agent-based simulation is the main method, though the application is transport and public health.", "model": "claude-sonnet-5-5"}
++{"work_id": 2968942120, "categories": ["database_systems"], "primary": "database_systems", "confidence": 0.6, "rationale": "Framework for publishing linked open data from heterogeneous sources using Semantic Web technologies, which is a data integration problem.", "model": "claude-sonnet-5-5"}
++{"work_id": 7130252400, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Mixed-methods management study of AI adoption in small business marketing. It is not computer-science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2911791709, "categories": [], "primary": null, "confidence": 0.7, "rationale": "E-learning course development and platform comparison. This is educational and not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4412987259, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control-theory consensus control of multi-agent systems with an observer. This is control engineering, not the AI agents category.", "model": "claude-sonnet-5-5"}
++{"work_id": 4289360054, "categories": ["ai_agents", "algorithms"], "primary": "ai_agents", "confidence": 0.5, "rationale": "Game-playing agent using minimax with alpha-beta pruning and an evaluation function with weight analysis.", "model": "claude-sonnet-5-5"}
++{"work_id": 3128896073, "categories": [], "primary": null, "confidence": 0.8, "rationale": "Information systems literature-review and theory-building methodology. It is not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 3037034904, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Transportation policy analysis using an agent-based simulation. The contribution is in transport engineering, not CS.", "model": "claude-sonnet-5-5"}
++{"work_id": 3014047513, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Algebraic controllability conditions for multiagent systems in control theory, which is outside the ontology.", "model": "claude-sonnet-5-5"}
++{"work_id": 3122776204, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Containment control of nonlinear multi-agent systems, which is control theory and not CS agents.", "model": "claude-sonnet-5-5"}
++{"work_id": 7077881861, "categories": ["reinforcement_learning", "ai_safety"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "Multi-agent RL method for mixed-motive cooperation with fairness guarantees; fairness is a secondary theme.", "model": "claude-sonnet-5-5"}
++{"work_id": 7160854959, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control-theory paper on event-triggered control of nonlinear multi-agent systems; not squarely within CS categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 3126546068, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Multi-agent system framework for power distribution network protection and restoration; agent-based but mainly power-engineering.", "model": "claude-sonnet-5-5"}
++{"work_id": 2997077628, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.65, "rationale": "Alternatives ranking for automated negotiation among multi-agent systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 3035359449, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.4, "rationale": "Uses multiagent methods to analyze EEG for brain activity recognition; the agent paradigm is the main method.", "model": "claude-sonnet-5-5"}
++{"work_id": 4212806673, "categories": [], "primary": null, "confidence": 0.7, "rationale": "Physics/biology thesis on reaction-diffusion pattern formation in Dictyostelium; not CS research.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416750122, "categories": ["reinforcement_learning", "robotics"], "primary": "reinforcement_learning", "confidence": 0.8, "rationale": "Decentralized multi-agent RL for large-scale mixed traffic control with robot vehicles, published at IROS.", "model": "claude-sonnet-5-5"}
++{"work_id": 2945626893, "categories": ["ai_agents", "ai_safety"], "primary": "ai_agents", "confidence": 0.5, "rationale": "Generative policy model for autonomous systems in coalitions using explainable AI; agent policy learning.", "model": "claude-sonnet-5-5"}
++{"work_id": 2969980661, "categories": ["distributed_computing", "computer_networks"], "primary": "distributed_computing", "confidence": 0.5, "rationale": "Diffusion/consensus protocol for networked multi-agent systems with convergence analysis and link failure detection.", "model": "claude-sonnet-5-5"}
++{"work_id": 4402920523, "categories": [], "primary": null, "confidence": 0.55, "rationale": "Control-theoretic consensus tracking for singular multi-agent systems; not CS-category specific.", "model": "claude-sonnet-5-5"}
++{"work_id": 7137881325, "categories": ["llm", "hci"], "primary": "hci", "confidence": 0.35, "rationale": "Essay on using AI and prompt engineering to teach ENEM essay writing; an educational application with weak technical contribution.", "model": "claude-sonnet-5-5"}
++{"work_id": 7218192222, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.3, "rationale": "Uses NetLogo multi-agent simulations as a teaching tool for complex systems; mostly education, with a weak link to multi-agent modelling.", "model": "claude-sonnet-5-5"}
++{"work_id": 2898995053, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Literature study of game characteristics in electronic negotiation support systems; information systems and social interaction, gamification.", "model": "claude-sonnet-5-5"}
++{"work_id": 3037046004, "categories": ["computer_vision", "hci"], "primary": "computer_vision", "confidence": 0.6, "rationale": "Collaborative satellite image analysis tool with human-in-the-loop retraining of neural networks for humanitarian use.", "model": "claude-sonnet-5-5"}
++{"work_id": 4416453808, "categories": ["ai_agents", "llm"], "primary": "ai_agents", "confidence": 0.92, "rationale": "Framework for designing LLM-based multi-agent systems with task decomposition, evaluated on knapsack and task-assignment problems.", "model": "claude-sonnet-5-5"}
++{"work_id": 4362606359, "categories": ["ml_theory", "ai_agents"], "primary": "ml_theory", "confidence": 0.6, "rationale": "Formal framework of structural causal games combining causal inference and game theory with multi-agent influence diagrams.", "model": "claude-sonnet-5-5"}
++{"work_id": 3004892764, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Interview study of conversational agents' effects on customers and employees via technology affordances; human-AI and social aspects.", "model": "claude-sonnet-5-5"}
++{"work_id": 4390619203, "categories": [], "primary": null, "confidence": 0.9, "rationale": "Analysis of the growth of distance education in Brazil; not computer science research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2898910837, "categories": ["ai_agents"], "primary": "ai_agents", "confidence": 0.3, "rationale": "Adaptive fuzzy consensus control for nonlinear multiagent systems; control theory that only loosely fits multi-agent systems.", "model": "claude-sonnet-5-5"}
++{"work_id": 7219983900, "categories": [], "primary": null, "confidence": 0.85, "rationale": "Information science and political economy study of the Brazilian land information regime; not computer science.", "model": "claude-sonnet-5-5"}
++{"work_id": 7132739348, "categories": ["formal_verification"], "primary": "formal_verification", "confidence": 0.5, "rationale": "Logic expressivity for KLM defeasible reasoning; knowledge representation/logic, closest to formal logic/automated reasoning.", "model": "claude-sonnet-5-5"}
++{"work_id": 7162579366, "categories": ["autonomous_scientific_discovery", "ai_agents"], "primary": "autonomous_scientific_discovery", "confidence": 0.92, "rationale": "End-to-end autonomous research system with verifiability audit for AI-generated research.", "model": "claude-sonnet-5-5"}
++{"work_id": 2920799053, "categories": ["algorithms"], "primary": "algorithms", "confidence": 0.35, "rationale": "Game-playing programs for Hex and Go; history and analysis of game AI, loosely search algorithms.", "model": "claude-sonnet-5-5"}
++{"work_id": 7164056265, "categories": ["ai_agents", "llm", "software_engineering"], "primary": "ai_agents", "confidence": 0.85, "rationale": "Design patterns for multi-agent LLM systems with tool agents in engineering case study.", "model": "claude-sonnet-5-5"}
++{"work_id": 7153320825, "categories": ["ai_agents", "generative_ai"], "primary": "ai_agents", "confidence": 0.7, "rationale": "Multi-agent system pipeline for documentary video production.", "model": "claude-sonnet-5-5"}
++{"work_id": 4415035946, "categories": ["ai_agents", "llm", "computational_biology"], "primary": "ai_agents", "confidence": 0.6, "rationale": "Cooperative LLM agents (Selector, Validator) for quantum chemistry property prediction.", "model": "claude-sonnet-5-5"}
++{"work_id": 2993634112, "categories": ["hci"], "primary": "hci", "confidence": 0.4, "rationale": "Educational study on mobile learning apps and computational thinking with students; user-centered design research.", "model": "claude-sonnet-5-5"}
++{"work_id": 7154382442, "categories": ["reinforcement_learning", "ai_agents", "llm"], "primary": "reinforcement_learning", "confidence": 0.85, "rationale": "MARL training of LLM agents with reward design for social deduction games.", "model": "claude-sonnet-5-5"}
++{"work_id": 7214219730, "categories": [], "primary": null, "confidence": 0.6, "rationale": "Control-theoretic consensus of linear multi-agent systems, systems and control rather than CS ontology categories.", "model": "claude-sonnet-5-5"}
++{"work_id": 7124257480, "categories": ["reinforcement_learning"], "primary": "reinforcement_learning", "confidence": 0.95, "rationale": "Cooperative MARL experience replay and reward decomposition method.", "model": "claude-sonnet-5-5"}

added taxonomy/validation/usage_claude-haiku-5-5.json +5 −0

++{
++ "in": 2394822,
++ "out": 192005,
++ "calls": 220
++}
+ \ No newline at end of file

added taxonomy/validation/usage_claude-sonnet-5-5.json +5 −0

++{
++ "in": 2383473,
++ "out": 178747,
++ "calls": 219
++}
+ \ No newline at end of file

spb/wp44_uqo

WP44: Part V (Agent E, 98 refs), Part IX results (21 tables), taxonomy validation metrics, KG extraction (11,086 works, 6.16 USD), A-008; work paused by the PI 2026-11-03 ~21:00

Theme