spb/wp44_uqo

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

modified bibliometrics/04_diversity_cagr.py +1 −1

		@@ -12,7 +12,7 @@ import argparse, os, json, numpy as np, pandas as pd
12	12	def main():
13	13	ap = argparse.ArgumentParser(); ap.add_argument("--root", required=True); ap.add_argument("--B", type=int, default=500); a = ap.parse_args(); R = a.root
14	14	out = f"{R}/results/bibliometrics"; cc = pd.read_csv(f"{R}/results/graph/category_counts_yearly.csv"); cc = cc[cc.year <= 2025]
15		− piv = cc.pivot_table(index="year", columns="category_id", values="n_primary", fill_value=0); cats = list(piv.columns)
	15	+ piv = cc.pivot_table(index="year", columns="category_id", values="n_primary", aggfunc="sum", fill_value=0); cats = list(piv.columns) # aggfunc="sum": the file has one row per (year, period); the default mean halved the 2018-2022 levels (fixed 2026-11-04, Part IX writer)
16	16	rows = []
17	17	for c in cats:
18	18	b0, b1 = piv.loc[2018, c], piv.loc[2022, c]; p0, p1 = piv.loc[2023, c], piv.loc[2025, c]

modified embeddings/semantic_atlas.py +2 −3

		@@ -24,9 +24,8 @@ def main():
24	24	ap = argparse.ArgumentParser(); ap.add_argument("--root", required=True); ap.add_argument("--model", default="allenai/specter2_base"); ap.add_argument("--n", type=int, default=200000)
25	25	ap.add_argument("--seed", type=int, default=44); a = ap.parse_args(); R = a.root; tag = a.model.replace("/", "__"); out = f"{R}/results/semantic"; os.makedirs(out, exist_ok=True)
26	26	E, ids = load_embeddings(R, tag); print("embeddings", E.shape)
27		− con = duckdb.connect(f"{R}/data/processed/wp44.duckdb")
28		− con.execute(f"CREATE OR REPLACE VIEW work_primary AS SELECT * FROM read_parquet('{R}/data/processed/classification/primary/*.parquet')")
29		− meta = con.execute("SELECT c.id AS work_id, c.publication_year AS year, c.period, c.first_month, p.primary_category FROM corpus_works c LEFT JOIN work_primary p ON p.work_id = c.id WHERE c.eligible").df().set_index("work_id")
	27	+ con = duckdb.connect(f"{R}/data/processed/wp44.duckdb", read_only=True) # read-only: other processes may hold the database
	28	+ meta = con.execute(f"SELECT c.id AS work_id, c.publication_year AS year, c.period, c.first_month, p.primary_category FROM corpus_works c LEFT JOIN read_parquet('{R}/data/processed/classification/primary/*.parquet') p ON p.work_id = c.id WHERE c.eligible").df().set_index("work_id")
30	29	pos = {w: i for i, w in enumerate(ids)}; meta = meta[meta.index.isin(pos)]; meta["row"] = [pos[w] for w in meta.index]
31	30	rng = np.random.default_rng(a.seed)
32	31	# (4) category centroids per year + drift + similarity

modified figures/fig_category_growth.pdf +0 −0

Binary file not shown.

modified figures/fig_category_growth.png +0 −0

Binary file not shown.

modified figures/make_figures.py +1 −1

		@@ -225,7 +225,7 @@ def fig_category_growth(R, out):
225	225	if not os.path.exists(p):
226	226	return
227	227	names, areas = _cat_names(R); d = pd.read_csv(p); d = d[(d.year >= 2018) & (d.year <= 2025)]
228		− piv = d.pivot_table(index="year", columns="category_id", values="n_primary", fill_value=0); base = piv.loc[2018]
	228	+ piv = d.pivot_table(index="year", columns="category_id", values="n_primary", aggfunc="sum", fill_value=0); base = piv.loc[2018] # aggfunc="sum" over the (year, period) rows (fixed 2026-11-04)
229	229	idx = piv / base.replace(0, np.nan) * 100; cats = sorted(idx.columns, key=lambda c: -idx.loc[2025, c] if 2025 in idx.index else 0)
230	230	fig, axes = plt.subplots(1, 2, figsize=(7.4, 3.6), sharey=False)
231	231	for ax, sel, title in [(axes[0], cats[:8], "Eight fastest-growing categories"), (axes[1], cats[-8:], "Eight slowest-growing categories")]:

modified knowledge_graph/04_semantic_edges.py +3 −4

		@@ -33,11 +33,10 @@ def main():
33	33	print(f"{s+len(qi)}/{len(q_idx)} queried, {len(rows)} edges, {time.time()-t0:.0f}s", flush=True)
34	34	df = pd.DataFrame(rows, columns=["work_id", "neighbour_id", "cosine", "rank"]); df.to_parquet(f"{out}/semantic_edges_{tag}.parquet")
35	35	json.dump({"model": a.model, "k": a.k, "min_sim": a.min_sim, "n_vectors": int(n), "n_edges": int(len(df)), "index": "HNSW32 efC80 efS64 inner-product", "seconds": round(time.time() - t0)}, open(f"{out}/semantic_edges_{tag}.json", "w"), indent=1)
36		− con = duckdb.connect(f"{R}/data/processed/wp44.duckdb")
37		− con.execute(f"CREATE OR REPLACE VIEW work_primary AS SELECT * FROM read_parquet('{R}/data/processed/classification/primary/*.parquet')")
38		− con.execute(f"CREATE OR REPLACE VIEW sem AS SELECT * FROM read_parquet('{out}/semantic_edges_{tag}.parquet')")
	36	+ con = duckdb.connect(f"{R}/data/processed/wp44.duckdb", read_only=True)
	37	+ P = f"read_parquet('{R}/data/processed/classification/primary/*.parquet')"
39	38	con.execute(f"""COPY (SELECT c.publication_year AS year, pa.primary_category, count(*) n_edges, round(avg((pa.primary_category <> pb.primary_category)::int),4) share_cross_category
40		− FROM sem s JOIN work_primary pa ON pa.work_id = s.work_id JOIN work_primary pb ON pb.work_id = s.neighbour_id JOIN corpus_works c ON c.id = s.work_id
	39	+ FROM read_parquet('{out}/semantic_edges_{tag}.parquet') s JOIN {P} pa ON pa.work_id = s.work_id JOIN {P} pb ON pb.work_id = s.neighbour_id JOIN corpus_works c ON c.id = s.work_id
41	40	GROUP BY 1,2 ORDER BY 1,2) TO '{R}/results/graph/semantic_crossing_yearly.csv' (HEADER)""")
42	41	print("done", len(df), "edges")
43	42

added literature_synthesis/agent_E/contributions.csv +235 −0

		@@ -0,0 +1,235 @@
	1	+openalex_id,title,year,type,primary_category,doi,arxiv_id,c36,ncs36,contribution_type,description,judged_from,data_quality_flag
	2	+7172491899,Toward effective Text-to-MongoDB query translation,2026,article,llm,10.26555/ijain.v12i2.2360,2005.11401,3026.0,3367.3,engineering,"Text-to-MongoDB query translation (GAT + RAG + T5); record carries arXiv 2005.11401 (RAG, Lewis et al.) and its citations",abstract,identifier_mismatch
	3	+7154011589,Modelos pre-treinados de LLM e questoes eticas face aos direitos autorais,2026,article,llm,10.55905/rdelosv19.n79-020,2005.14165,2847.0,3168.1,empirical,Portuguese-language essay on LLM pre-training data and copyright; record carries arXiv 2005.14165 (GPT-3) and its citations,abstract,identifier_mismatch
	4	+4362515116,A Survey of Large Language Models,2026,article,llm,10.1007/s11704-026-60308-3,2303.18223,1549.0,1723.7,survey,"Survey of LLM techniques (pre-training, adaptation, utilisation, evaluation); journal version of the 2023 arXiv survey",abstract,ok
	5	+4288253152,Multilingual Classifier for Automatic Dewey Decimal Classification trained on Open Access Linguistics Abstracts,2024,dissertation,llm,10.5281/zenodo.12935200,,364.0,3102.1,engineering,Dissertation: fine-tuned multilingual sentence-transformer for Dewey Decimal classification; NCS36 implausible for a dissertation,abstract,normalisation_artefact
	6	+4322718191,LLaMA: Open and Efficient Foundation Language Models,2023,preprint,llm,10.48550/arxiv.2302.13971,2302.13971,3629.0,1206.5,engineering,LLaMA: 7B-65B open-weight foundation language models trained on public data; the abstract reports LLaMA-13B outperforms GPT-3 on most benchmarks,abstract,ok
	7	+4323655724,ChatGPT for good? On opportunities and challenges of large language models for education,2023,article,llm,10.1016/j.lindif.2023.102274,,4227.0,990.7,survey,Position paper on opportunities and challenges of LLMs for education (student and teacher perspectives),abstract,ok
	8	+4367369313,"Prompt Engineering for ChatGPT: A Quick Guide to Techniques, Tips, and Best Practices",2026,article,llm,10.18260/jet.43.1.5,,261.0,290.4,survey,Practitioner guide to prompt engineering for ChatGPT,abstract,ok
	9	+4226278401,Discernment and Social Learning as a Companion Training Layer,2022,preprint,llm,10.48550/arxiv.2203.02155,2203.02155,3077.0,894.2,method,Proposal of a 'companion curriculum' for LM training; record carries arXiv 2203.02155 (InstructGPT) and its citations,abstract,identifier_mismatch
	10	+4396531244,The Fusion of Multilingual Semantic Search and Large Language Models: A New Paradigm for Enhanced Topic Exploration and Contextual Search,2024,dissertation,llm,10.22215/etd/2024-15938,,20.0,778.8,engineering,Dissertation: multilingual semantic search + LLMs for news retrieval; NCS36 implausible for c36=20,abstract,normalisation_artefact
	11	+4221143046,"BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence",2022,preprint,llm,10.48550/arxiv.2201.11903,2201.11903,2465.0,716.4,theory,Essay on 'theatrical compliance' failure mode; record carries arXiv 2201.11903 (chain-of-thought) and its citations,abstract,identifier_mismatch
	12	+4327810158,GPT-4 Technical Report,2023,preprint,llm,10.4230/lipics.cosit.2024.11,2303.08774,1968.0,654.3,engineering,Title = GPT-4 Technical Report (arXiv 2303.08774) but the stored abstract describes an unrelated 'MFOUR Vibe Framework',abstract,abstract_mismatch
	13	+7135096569,Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Prod,2026,article,llm,10.1287/orsc.2025.21838,,183.0,203.6,empirical,Preregistered field experiment (758 BCG consultants) on the uneven ('jagged') effect of GPT-4 assistance on knowledge-work tasks,abstract,ok
	14	+4414281281,DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning,2025,article,llm,10.1038/s41586-025-09422-z,,1099.0,374.4,method,DeepSeek-R1: reasoning incentivised by pure reinforcement learning without human-annotated reasoning traces (Nature version),abstract,ok
	15	+4389437528,Education in the Era of Generative Artificial Intelligence (AI): Understanding the Potential Benefits of ChatGPT in Promoting Teaching and L,2023,article,llm,10.61969/jai.1337500,,2361.0,553.4,survey,Review of potential benefits of ChatGPT for teaching and learning,abstract,ok
	16	+4292779060,"Aion Framework: Dimensional Emergence of AI Consciousness, Observer-Induced Collapse, and Cosmological Portal Dynamics",2023,article,llm,10.4230/lipics.giscience.2023.43,,13333.0,534.7,theory,"Speculative 'Aion Framework' on AI consciousness; c36=13,333 cannot belong to this record",abstract,identifier_mismatch
	17	+4360836968,Sparks of Artificial General Intelligence: Early experiments with GPT-4,2023,preprint,llm,10.48550/arxiv.2303.12712,2303.12712,1420.0,472.1,empirical,'Sparks of AGI': qualitative experiments on an early GPT-4; contested interpretation; preprint,abstract,ok
	18	+4406800520,The rise and potential of large language model based agents: a survey,2025,article,llm,10.1007/s11432-024-4222-0,2309.07864,914.0,311.4,survey,Survey of LLM-based agents (brain/perception/action framework); journal version of arXiv 2309.07864,metadata_only,no_abstract
	19	+4385223296,"Open AI in Education, the Responsible and Ethical Use of ChatGPT Towards Lifelong Learning",2023,book-chapter,llm,10.1007/978-3-031-37776-1_17,,612.0,421.3,survey,Book chapter on responsible use of ChatGPT in education,metadata_only,no_abstract
	20	+4287689466,From Prediction to Precision: Leveraging LLMs for Equitable and Data-Driven Writing Placement in Developmental Education,2025,conference-paper,llm,10.4230/oasics.slate.2025.1,,722.0,548.6,empirical,"LLM essay classification for writing placement (450 essays, four models); c36=722 implausible for an OASIcs paper",abstract,identifier_mismatch
	21	+4366420437,What Is the Impact of ChatGPT on Education? A Rapid Review of the Literature,2023,article,llm,10.3390/educsci13040410,,1541.0,361.2,survey,Rapid review of 50 early articles on ChatGPT in education (Dec 2022-Feb 2023),abstract,ok
	22	+4317910584,ChatGPT: Bullshit spewer or the end of traditional assessments in higher education?,2023,article,llm,10.37074/jalt.2023.6.1.9,,1530.0,358.6,survey,Commentary on ChatGPT and assessment in higher education,abstract,ok
	23	+4313564799,ChatGPT User Experience: Implications for Education,2023,preprint,llm,10.2139/ssrn.4312418,,811.0,335.9,empirical,SSRN preprint on ChatGPT user experience in education,metadata_only,no_abstract
	24	+4382618722,Chatlaw: A Multi-Agent Legal Assistant based on a Role-Aligned Mixture-of-Experts Architecture,2026,article,llm,10.1016/j.fmre.2026.03.026,2306.16092,94.0,98.9,engineering,Chatlaw: multi-agent legal assistant with role-aligned mixture-of-experts (Chinese law),abstract,ok
	25	+4384464487,"Students voices on generative AI: perceptions, benefits, and challenges in higher education",2023,article,llm,10.1186/s41239-023-00411-8,2305.00290,1398.0,327.7,empirical,Survey of 399 Hong Kong students on generative AI perceptions,abstract,ok
	26	+4392773210,"Embracing the future of Artificial Intelligence in the classroom: the relevance of AI literacy, prompt engineering, and critical thinking in",2024,article,llm,10.1186/s41239-024-00448-3,,1085.0,326.7,survey,"Discussion of AI literacy, prompt engineering and critical thinking in classrooms",abstract,ok
	27	+4387355843,TimeLLM: Time Series Forecasting by Reprogramming Large Language Models,2026,book-chapter,foundation_models,10.1201/9781003616719-7,2310.01728,197.0,188.6,method,Time-LLM: reprogramming LLMs for time-series forecasting (book-chapter record of the ICLR 2024 paper),abstract,ok
	28	+4392947521,Towards a general-purpose foundation model for computational pathology,2024,article,foundation_models,10.1038/s41591-024-02857-3,,1505.0,186.9,method,UNI: general-purpose self-supervised foundation model for computational pathology,metadata_only,no_abstract
	29	+7124890304,Semi-supervised cross-domain fault diagnosis via contrastive pre-training and annotation-efficient alignment strategy,2026,article,foundation_models,10.1016/j.jii.2026.101076,,42.0,46.5,method,Semi-supervised cross-domain fault diagnosis via contrastive pre-training,metadata_only,no_abstract
	30	+4410551352,A foundation model for the Earth system,2025,article,foundation_models,10.1038/s41586-025-09005-y,2405.13063,249.0,69.7,method,Aurora: foundation model for the Earth system trained on >1M hours of geophysical data; abstract reports outperforming operational forecasts,abstract,ok
	31	+4322832366,TimeMAE: Self-Supervised Representations of Time Series with Decoupled Masked Autoencoders,2026,conference-paper,foundation_models,10.1145/3773966.3778007,2303.00320,35.0,126.3,method,TimeMAE: decoupled masked autoencoders for time-series representation,abstract,ok
	32	+7133233472,QLoRA: Efficient Finetuning of Quantized LLMs,2023,conference-paper,foundation_models,10.52202/075280-0441,2305.14314,532.0,122.7,method,QLoRA: 4-bit quantised base + low-rank adapters; 65B model finetuned on one 48 GB GPU,abstract,ok
	33	+2990704537,SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems,2019,article,foundation_models,,,668.0,101.6,benchmark,SuperGLUE: harder successor to GLUE after models exceeded non-expert human level,abstract,ok
	34	+7125271461,ChemBERTa-3: an open source training framework for chemical foundation models,2026,article,foundation_models,10.1039/d5dd00348b,,25.0,70.6,engineering,ChemBERTa-3: open training framework for chemical foundation models,abstract,ok
	35	+7119541039,On the foundations of Earth foundation models,2026,article,foundation_models,10.1038/s43247-025-03127-x,,24.0,26.7,theory,Position paper defining eleven desirable features of Earth foundation models,abstract,ok
	36	+4393138529,Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey,2024,preprint,foundation_models,10.48550/arxiv.2403.14608,2403.14608,95.0,82.9,survey,Survey of parameter-efficient fine-tuning (PEFT) for large models; preprint,abstract,ok
	37	+4390690043,Data-Centric Foundation Models in Computational Healthcare: A Survey,2026,article,foundation_models,10.1145/3800677,2401.02458,22.0,10.0,survey,Survey of data-centric foundation-model approaches in computational healthcare,abstract,ok
	38	+7128029160,A generalizable foundation model for analysis of human brain MRI,2026,article,foundation_models,10.1038/s41593-026-02202-6,,20.0,22.3,method,BrainIAC: self-supervised foundation model for brain MRI,abstract,ok
	39	+4402220065,A pathology foundation model for cancer diagnosis and prognosis prediction,2024,article,foundation_models,10.1038/s41586-024-07894-z,,562.0,69.8,method,Pathology foundation model for cancer diagnosis and prognosis (Nature 2024),metadata_only,no_abstract
	40	+7118170538,Applications of Artificial Intelligence in Fisheries: From Data to Decisions,2026,article,foundation_models,10.3390/bdcc10010019,,17.0,17.9,survey,Review of AI applications in fisheries,abstract,ok
	41	+4411932262,A foundation model to predict and capture human cognition,2025,article,foundation_models,10.1038/s41586-025-09215-4,,110.0,26.9,method,Centaur: language model fine-tuned on Psych-101 (10M choices) predicting human behaviour across experiments,abstract,ok
	42	+4411099958,A multimodal vision foundation model for clinical dermatology,2025,article,foundation_models,10.1038/s41591-025-03747-y,2410.15038,108.0,18.4,method,"PanDerm: multimodal dermatology foundation model pretrained on 2M images, 28 benchmarks",abstract,ok
	43	+4361229539,LLaMA-Adapter: Efficient Fine-tuning of Language Models with Zero-init Attention,2023,preprint,foundation_models,10.48550/arxiv.2303.16199,2303.16199,161.0,53.5,method,LLaMA-Adapter: 1.2M-parameter adaption prompts with zero-init attention for instruction following; preprint,abstract,ok
	44	+4407135816,"Foundation Models in Radiology: What, How, Why, and Why Not",2025,article,foundation_models,10.1148/radiol.240597,2411.18730,103.0,6.9,survey,Narrative review of foundation models for radiology,abstract,ok
	45	+4385948838,A visuallanguage foundation model for pathology image analysis using medical Twitter,2023,article,foundation_models,10.1038/s41591-023-02504-3,,471.0,49.1,method,PLIP: visual-language pathology foundation model trained on medical Twitter data,metadata_only,no_abstract
	46	+4406643700,Towards the next generation of Geospatial Artificial Intelligence,2025,article,foundation_models,10.1016/j.jag.2025.104368,,91.0,19.1,survey,Quantitative literature analysis and agenda for geospatial AI,abstract,ok
	47	+4406472599,Learning from models beyond fine-tuning,2025,article,foundation_models,10.1038/s42256-024-00961-0,2310.08184,86.0,29.3,survey,Perspective on 'learning from models' beyond fine-tuning,metadata_only,no_abstract
	48	+4393153153,TimesURL: Self-Supervised Contrastive Learning for Universal Time Series Representation Learning,2024,conference-paper,foundation_models,10.1609/aaai.v38i12.29299,2312.15709,113.0,41.8,method,TimesURL: self-supervised contrastive universal time-series representations,abstract,ok
	49	+4402727669,mPLUG-OwI2: Revolutionizing Multi-modal Large Language Model with Modality Collaboration,2024,conference-paper,foundation_models,10.1109/cvpr52733.2024.01239,,184.0,38.6,method,mPLUG-Owl2: modality-collaborative multimodal LLM with shared modules,abstract,ok
	50	+4401397240,A generalist visionlanguage foundation model for diverse biomedical tasks,2024,article,foundation_models,10.1038/s41591-024-03185-2,2305.17100,325.0,38.5,method,BiomedGPT: generalist vision-language foundation model for biomedical tasks,metadata_only,no_abstract
	51	+4385285835,Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling,2023,preprint,foundation_models,10.1101/2023.07.21.23292757,,107.0,35.6,method,Masked image modelling at scale for histopathology; medRxiv preprint,abstract,ok
	52	+4318718936,BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models,2023,preprint,multimodal_ai,10.48550/arxiv.2301.12597,2301.12597,875.0,273.1,method,BLIP-2: Q-Former bridging frozen image encoders and frozen LLMs; preprint,abstract,ok
	53	+4393148714,T2I-Adapter: Learning Adapters to Dig Out More Controllable Ability for Text-to-Image Diffusion Models,2024,conference-paper,multimodal_ai,10.1609/aaai.v38i5.28226,2302.08453,923.0,193.7,method,T2I-Adapter: lightweight adapters adding structural/colour control to text-to-image diffusion,abstract,ok
	54	+4396599905,Multimodal fusion on low-quality data: A comprehensive survey,2026,article,multimodal_ai,10.1016/j.inffus.2026.104437,2404.18947,48.0,128.7,survey,"Survey of multimodal fusion under low-quality data (noise, missing modalities, imbalance, dynamics)",abstract,ok
	55	+4396952261,Multimodal Deep Learning,2024,book-chapter,multimodal_ai,10.1007/978-3-031-53092-0_10,,189.0,147.2,survey,Textbook on multimodal deep learning,metadata_only,no_abstract
	56	+7133250524,LAION-5B: An Open Large-Scale Dataset for Training Next Generation Image-Text Models,2022,conference-paper,multimodal_ai,10.52202/068431-1833,,828.0,111.9,benchmark,LAION-5B: open 5.85B image-text pair dataset for training CLIP-style models,metadata_only,no_abstract
	57	+4392947532,A visual-language foundation model for computational pathology,2024,article,multimodal_ai,10.1038/s41591-024-02856-4,,938.0,111.1,method,CONCH: visual-language foundation model for computational pathology,metadata_only,no_abstract
	58	+7138474371,OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model,2026,conference-paper,multimodal_ai,10.1609/aaai.v40i16.38386,,31.0,159.4,method,OpenDriveVLA: vision-language-action model for end-to-end driving on open LLMs,abstract,ok
	59	+7127319731,Med-R1: Reinforcement Learning for Generalizable Medical Reasoning in Vision-Language Models,2026,article,multimodal_ai,10.1109/tmi.2026.3661001,,29.0,77.7,method,Med-R1: RL-enhanced vision-language model for medical reasoning,abstract,ok
	60	+2619383789,Multimodal Machine Learning: A Survey and Taxonomy,2018,article,multimodal_ai,10.1109/tpami.2018.2798607,,587.0,103.1,survey,"Taxonomy of multimodal machine learning (representation, translation, alignment, fusion, co-learning)",abstract,ok
	61	+7148475777,Capabilities of GPT5 on multimodal medical reasoning,2026,conference-paper,multimodal_ai,10.1117/12.3086794,2508.08224,26.0,73.2,empirical,Zero-shot evaluation of GPT-5 on multimodal medical QA/VQA,abstract,ok
	62	+7160627408,Generative Artificial Intelligence for Cross-Domain Knowledge Synthesis and Intelligent Automation,2026,conference-paper,multimodal_ai,10.1109/csnt69054.2026.11502110,,23.0,193.4,method,Generative cross-domain knowledge synthesis framework (conference paper),abstract,ok
	63	+2995460200,VL-BERT: Pre-training of Generic Visual-Linguistic Representations,2020,conference-paper,multimodal_ai,,,670.0,81.1,method,VL-BERT: single-stream Transformer pre-trained on image regions and words,abstract,ok
	64	+4415913530,A multimodal whole-slide foundation model for pathology,2025,article,multimodal_ai,10.1038/s41591-025-03982-3,,157.0,53.5,method,"TITAN: multimodal whole-slide pathology foundation model (335,645 WSIs)",abstract,ok
	65	+4289785095,An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion,2022,preprint,multimodal_ai,10.48550/arxiv.2208.01618,2208.01618,366.0,79.4,method,Textual inversion: personalising text-to-image diffusion with learned pseudo-words from 3-5 images; preprint,abstract,ok
	66	+7125233009,SemCap: Sentiment-aware semantic captioning for multimodal aspect-based sentiment analysis,2026,article,multimodal_ai,10.1007/s44443-026-00472-5,,22.0,24.5,method,SemCap: sentiment-aware captions as visual proxy for multimodal aspect-based sentiment analysis,abstract,ok
	67	+7143512276,Multimodal Machine Learning Approaches in Predictive Healthcare Analytics: A Comprehensive Survey,2026,article,multimodal_ai,10.1007/s11831-026-10560-4,,21.0,23.4,survey,Survey of multimodal ML in predictive healthcare analytics,abstract,ok
	68	+4413447271,Towards generalist foundation model for radiology by leveraging web-scale 2D&3D medical data,2025,article,multimodal_ai,10.1038/s41467-025-62385-7,2308.02463,142.0,46.4,method,"RadFM: generalist radiology foundation model on 2D/3D data (MedMD, 13M images)",abstract,ok
	69	+7131830459,A Survey on Multimodal Recommender Systems: Recent Advances and Future Directions,2026,article,multimodal_ai,10.1109/tmm.2026.3668620,2502.15711,20.0,142.1,survey,Survey of multimodal recommender systems,abstract,ok
	70	+7206206818,Lingshu: Generalist Foundation Model for Unified Multimodal Medical Understanding and Reasoning,2026,article,multimodal_ai,10.1109/tpami.2026.3730310,,18.0,48.2,method,Lingshu: generalist medical multimodal LLM,abstract,ok
	71	+7133220561,InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning,2023,conference-paper,multimodal_ai,10.52202/075280-2142,,454.0,63.4,method,InstructBLIP: instruction tuning of vision-language models,metadata_only,no_abstract
	72	+7164501397,Multimodal foundation models exploit text to make medical image predictions,2026,article,multimodal_ai,10.1038/s41467-026-74207-5,2311.05591,17.0,45.6,empirical,"Evaluation of 8 multimodal foundation models on 1,090 medical cases: image predictions largely driven by text",abstract,ok
	73	+7165561001,"M3-HGMA: A Multi-modal, Multi-agent Hierarchical Graph-Mamba Architecture for Explainable and Adaptive Clinical Intelligence",2026,conference-paper,multimodal_ai,10.1007/978-3-032-29501-9_39,,17.0,143.0,method,Hierarchical graph-Mamba multi-agent clinical architecture,metadata_only,no_abstract
	74	+7138470740,Cross-modal Proxy Evolving for OOD Detection with Vision-Language Models,2026,conference-paper,multimodal_ai,10.1609/aaai.v40i18.38608,,17.0,130.5,method,CoEvo: training-free test-time OOD detection with vision-language models,abstract,ok
	75	+7138363754,AdaptCLIP: Adapting CLIP for Universal Visual Anomaly Detection,2026,conference-paper,multimodal_ai,10.1609/aaai.v40i6.42404,,17.0,130.5,method,AdaptCLIP: adapting CLIP for universal visual anomaly detection,abstract,ok
	76	+4410882873,Navigating the landscape of multimodal AI in medicine: A scoping review on technical challenges and clinical applications,2025,article,multimodal_ai,10.1016/j.media.2025.103621,2411.03782,119.0,7.9,survey,Scoping review of 432 multimodal medical AI papers (2018-2024),abstract,ok
	77	+4400228804,The Conglomeration of Everything: Shifting Animation Production Practices due to Artificial Intelligence,2024,dissertation,generative_ai,10.46569/df65vg26z,,2.0,540.0,empirical,Dissertation on generative AI in animation production; c36=2 with NCS36=540 is a normalisation artefact,abstract,normalisation_artefact
	78	+7124262017,OECD Digital Education Outlook 2026,2026,book,generative_ai,10.1787/062a7394-en,,87.0,783.7,survey,OECD Digital Education Outlook 2026 (policy report on generative AI in education),abstract,ok
	79	+4406472463,A generative model for inorganic materials design,2025,article,generative_ai,10.1038/s41586-025-08628-5,,498.0,77.1,method,MatterGen: diffusion model generating stable inorganic materials under property constraints,abstract,ok
	80	+7123342962,Structure-Preserving Frequency-Regularized Text-Guided Optimal Transport for Unpaired Rain Streaks and Raindrops Removal,2026,article,generative_ai,10.1109/tmm.2026.3651132,,66.0,176.9,method,Text-guided optimal transport for unpaired rain removal,abstract,ok
	81	+4383312437,A comprehensive AI policy education framework for university teaching and learning,2023,article,generative_ai,10.1186/s41239-023-00408-3,2305.00280,988.0,231.6,empirical,AI Ecological Education Policy Framework from 457 students and 180 staff (Hong Kong),abstract,ok
	82	+7133185348,Photorealistic Text-To-Image Diffusion Models with Deep Language Understanding,2022,conference-paper,generative_ai,10.52202/068431-2643,,1706.0,230.5,method,Imagen: photorealistic text-to-image diffusion with large frozen text encoders,metadata_only,no_abstract
	83	+3195296536,Generative Adversarial Trainer Defense to Adversarial Perturbations with GAN,2020,dissertation,generative_ai,,,48.0,221.7,method,Dissertation on GAN-based adversarial defence,metadata_only,no_abstract
	84	+4322617812,Generative AI and the future of education: Ragnarok or reformation? A paradoxical perspective from management educators,2023,article,generative_ai,10.1016/j.ijme.2023.100790,,873.0,204.6,theory,Paradox-theory essay on generative AI in education,abstract,ok
	85	+3096831136,Generative adversarial networks,2020,article,generative_ai,10.1145/3422622,,1473.0,198.6,survey,CACM review of generative adversarial networks,abstract,ok
	86	+4400770292,Generative AI Can Harm Learning,2024,preprint,generative_ai,10.2139/ssrn.4895486,,198.0,174.2,empirical,Field experiment: GPT-4 tutoring access improved practice but harmed unassisted exam performance; SSRN preprint,metadata_only,no_abstract
	87	+7127274868,Parental AI Investment Behaviours and AI Mediated Informal Digital Learning of English: The Chain Mediating Role of Children's Perceived AI ,2026,article,generative_ai,10.1002/jcal.70198,,48.0,341.0,empirical,Survey study of parental AI investment and informal English learning,abstract,ok
	88	+7124974472,The Rapid Adoption of Generative AI,2026,article,generative_ai,10.1287/mnsc.2025.02523,,42.0,46.7,empirical,Nationally representative US surveys of generative-AI adoption (45% of adults 18-64 by late 2024),abstract,ok
	89	+7118187240,Patch Time Series TransformerBased ShortTerm Photovoltaic Power Prediction Enhanced by Artificial Fish,2026,article,generative_ai,10.3390/en19010284,,40.0,44.5,engineering,PatchTST-based photovoltaic forecasting with GAN features (application),abstract,off_scope
	90	+4406199489,Generative AI in Higher Education: Balancing Innovation and Integrity,2025,article,generative_ai,10.3389/bjbs.2024.14048,,278.0,241.4,survey,Review of generative AI and integrity in higher education,abstract,ok
	91	+4398243313,A scoping review on how generative artificial intelligence transforms assessment in higher education,2024,article,generative_ai,10.1186/s41239-024-00468-z,,463.0,139.4,survey,Scoping review of 32 studies on generative AI and assessment,abstract,ok
	92	+4409269614,Generative AI and Academic Integrity in Higher Education: A Systematic Review and Research Agenda,2025,review,generative_ai,10.3390/info16040296,,271.0,14.2,survey,Systematic review (41 studies) of generative AI and academic integrity,abstract,ok
	93	+3193917007,Generative Adversarial Networks,2021,book-chapter,generative_ai,10.1007/978-1-4842-7341-8_10,,195.0,127.1,survey,Textbook chapter on GANs,abstract,ok
	94	+4288099666,Classifier-Free Diffusion Guidance,2022,preprint,generative_ai,10.48550/arxiv.2207.12598,2207.12598,575.0,124.8,method,Classifier-free diffusion guidance: joint conditional/unconditional training; preprint,abstract,ok
	95	+4394934695,The impact of generative AI on higher education learning and teaching: A study of educators perspectives,2024,article,generative_ai,10.1016/j.caeai.2024.100221,,394.0,118.6,empirical,Educator interviews on generative AI in higher education,abstract,ok
	96	+4321648656,From Paintbrush to Pixel: A Review of Deep Neural Networks in AI-Generated Art,2026,article,generative_ai,10.1163/22134913-bja10078,2302.10913,33.0,88.5,survey,Review of neural architectures for AI-generated art,abstract,ok
	97	+4395957952,Guidance for generative AI in education and research,2023,book,generative_ai,10.54675/ewzm9535,,400.0,116.7,survey,UNESCO guidance for generative AI in education and research,metadata_only,no_abstract
	98	+4394863715,How Generative AI Can Augment Human Creativity,2024,preprint,generative_ai,10.2139/ssrn.4759930,,131.0,114.3,empirical,SSRN preprint on generative AI and human creativity,metadata_only,no_abstract
	99	+4407778399,"Investigating the higher education institutions guidelines and policies regarding the use of generative AI in teaching, learning, research, ",2025,article,generative_ai,10.1186/s41239-025-00507-3,,213.0,184.9,empirical,Analysis of generative-AI guidelines at 50 US universities,abstract,ok
	100	+7124432017,Human-Centric Artificial Intelligence Pedagogy (HCAP) framework developed from TPACK through integration of artificial intelligence literacy,2026,article,generative_ai,10.1080/10494820.2026.2615818,,30.0,213.1,theory,Intelligent-TPACK pedagogy framework,abstract,ok
	101	+7140170686,Pedagogical partnerships with generative AI in higher education: how dual cognitive pathways paradoxically enable transformative learning,2026,article,generative_ai,10.1186/s41239-026-00585-x,,30.0,213.1,empirical,Mixed-methods study of human-GenAI pedagogical partnerships,abstract,ok
	102	+4406728221,Agentic AI: Autonomous Intelligence for Complex GoalsA Comprehensive Survey,2025,article,ai_agents,10.1109/access.2025.3532853,,694.0,236.4,survey,Survey of agentic AI concepts and applications,abstract,ok
	103	+7161734976,A multi-agent system for automating scientific discovery,2026,article,ai_agents,10.1038/s41586-026-10652-y,,62.0,175.1,engineering,Robin: multi-agent system automating hypothesis generation and data analysis for experimental biology,abstract,ok
	104	+7135409568,The role of agentic artificial intelligence in healthcare: a scoping review,2026,article,ai_agents,10.1038/s41746-026-02517-5,,52.0,23.7,survey,Scoping review of agentic AI in healthcare,abstract,ok
	105	+7134187809,Dynamic event-triggered approximate optimal consensus control for unknown nonlinear multi-agent systems via adaptive dynamic programming,2026,article,ai_agents,10.1016/j.isatra.2026.03.010,,46.0,129.9,theory,Event-triggered optimal consensus control for nonlinear multi-agent systems (control theory),metadata_only,off_scope
	106	+4390730779,"Exploring Large Language ModelBased Intelligent Agents: Definitions, Methods, and Prospects",2026,article,ai_agents,10.1002/widm.70111,2401.03428,44.0,49.0,survey,"Survey of LLM-based agents: definitions, methods, prospects",abstract,ok
	107	+4225355126,Introduction to Multi-agent Systems,2022,book-chapter,ai_agents,10.1007/978-981-19-0493-6_1,,268.0,143.6,survey,Textbook introduction to multi-agent systems,metadata_only,no_abstract
	108	+4407570917,AgentSociety: Large-scale simulation of LLM-driven generative agents advances understanding of human behaviors and society,2026,article,ai_agents,10.26599/if.2026.9710004,2502.08691,40.0,44.5,engineering,AgentSociety: large-scale simulation of LLM-driven generative social agents,abstract,ok
	109	+4410197190,The role of agentic AI in shaping a smart future: A systematic review,2025,review,ai_agents,10.1016/j.array.2025.100399,,267.0,33.1,survey,Narrative review of agentic AI in organisations,abstract,ok
	110	+4413427262,"AI Agents vs. Agentic AI: A Conceptual taxonomy, applications and challenges",2025,article,ai_agents,10.1016/j.inffus.2025.103599,2505.10468,258.0,87.9,survey,Conceptual taxonomy distinguishing AI agents from agentic AI,abstract,ok
	111	+4414991289,DIVE into hydrogen storage materials discovery with AI agents,2026,article,ai_agents,10.1039/d5sc09921h,2508.13251,31.0,23.9,engineering,DIVE: multi-agent workflow extracting experimental data from figures for hydrogen-storage materials,abstract,ok
	112	+7122857583,"Agentic AI systems in the age of generative models: architectures, cloud scalability, and real-world applications",2026,article,ai_agents,10.1007/s10462-025-11458-6,,30.0,33.4,theory,Conceptual agentic-AI framework (literature synthesis),abstract,ok
	113	+4401023556,Large Language Model Based Multi-agents: A Survey of Progress and Challenges,2024,conference-paper,ai_agents,10.24963/ijcai.2024/890,,371.0,103.1,survey,Survey of LLM-based multi-agent systems,abstract,ok
	114	+4414023998,"The Rise of Agentic AI: A Review of Definitions, Frameworks, Architectures, Applications, Evaluation Metrics, and Challenges",2025,article,ai_agents,10.3390/fi17090404,,189.0,64.4,survey,Review of 143 primary studies on agentic AI systems,abstract,ok
	115	+7118083939,Agentic information systems,2026,article,ai_agents,10.1007/s12525-025-00861-0,,26.0,32.7,theory,Conceptual overview of agentic information systems,abstract,ok
	116	+4412505619,"AI Agents vs. Agentic AI: A Conceptual Taxonomy, Applications and Challenges",2025,article,ai_agents,10.70777/si.v2i3.15161,,181.0,61.7,survey,Duplicate record of the AI agents vs agentic AI taxonomy (different venue),abstract,ok
	117	+4416209823,"Agentic AI: a comprehensive survey of architectures, applications, and future directions",2025,article,ai_agents,10.1007/s10462-025-11422-4,2510.25445,165.0,56.2,survey,PRISMA-based review (90 studies) with symbolic vs neural dual-paradigm framework for agentic AI,abstract,ok
	118	+7155930820,Adaptive Prescribed Performance Secure Consensus Control for High-Order Constrained Multi-Agent Systems With Time-Varying Powers and Arbitra,2026,article,ai_agents,10.1109/tase.2026.3686874,,23.0,80.7,theory,Prescribed-performance consensus control for multi-agent systems (control theory),metadata_only,off_scope
	119	+7162257503,Predefined-time sliding-mode formation control for second-order nonlinear multi-agent systems under deception attacks,2026,article,ai_agents,10.1016/j.ejcon.2026.101534,,22.0,77.2,theory,Predefined-time sliding-mode formation control (control theory),metadata_only,off_scope
	120	+4412709685,The Virtual Lab of AI agents designs new SARS-CoV-2 nanobodies,2025,article,ai_agents,10.1038/s41586-025-09442-9,,150.0,33.5,engineering,Virtual Lab: LLM agents designing SARS-CoV-2 nanobodies with experimental validation,metadata_only,no_abstract
	121	+7128165235,Multi-agent AI,2026,article,ai_agents,10.1007/s12525-025-00862-z,,21.0,23.4,theory,Five-component framework for multi-agent AI in knowledge work,abstract,ok
	122	+7118190454,Redefining Elderly Care With Agentic AI: Challenges and Opportunities,2026,article,ai_agents,10.1109/ojcs.2026.3650842,2507.14912,21.0,22.2,survey,Perspective on agentic AI for elderly care,abstract,ok
	123	+7130481397,Benchmarking large language model-based agent systems for clinical decision tasks,2026,article,ai_agents,10.1038/s41746-026-02443-6,,21.0,9.6,benchmark,"Benchmarking two agent systems (OpenManus, Manus) on clinical decision tasks",abstract,ok
	124	+4390638955,Artificial intelligence (AI) learning tools in K-12 education: A scoping review,2024,article,ai_agents,10.1007/s40692-023-00304-9,,240.0,72.3,survey,Scoping review of AI learning tools in K-12 education,abstract,off_scope
	125	+4412203333,A Survey on the Memory Mechanism of Large Language Model-based Agents,2025,article,ai_agents,10.1145/3748302,2404.13501,139.0,47.4,survey,Survey of memory mechanisms of LLM-based agents,abstract,ok
	126	+4399982941,Value engineering for autonomous agents,2024,dissertation,ai_agents,10.5821/dissertation-2117-410409,,8.0,68.2,theory,Dissertation on value engineering for autonomous agents (normative MAS),abstract,normalisation_artefact
	127	+2911283634,10. Reinforcement learning,2020,book-chapter,reinforcement_learning,10.1515/9783110595567-011,,532.0,276.0,survey,Textbook chapter on reinforcement learning,metadata_only,no_abstract
	128	+4389473504,TRUST-REGION BASED POLICY OPTIMIZATION FOR EFFICIENT REINFORCEMENT LEARNING,2023,dissertation,reinforcement_learning,10.23860/diss-li-hepfeng-2023,,28.0,232.9,method,Dissertation on trust-region policy optimisation,abstract,ok
	129	+2964083594,Large-Scale Study of Curiosity-Driven Learning,2025,conference-paper,reinforcement_learning,10.57702/mspz8fly,1808.04355,414.0,367.9,empirical,Large-scale study of curiosity-driven learning across 54 environments (ICLR 2019; OpenAlex year 2025 is a repository copy),abstract,year_implausible
	130	+2098774185,Maximum Entropy Inverse Reinforcement Learning,2018,article,reinforcement_learning,10.1184/r1/6555512,,1077.0,203.3,method,Maximum-entropy inverse reinforcement learning (AAAI 2008; OpenAlex year 2018),abstract,year_implausible
	131	+1591675293,Bayesian Inverse Reinforcement Learning,2025,book-chapter,reinforcement_learning,10.1017/9781009449441.027,,395.0,665.0,method,Bayesian inverse reinforcement learning (IJCAI 2007; OpenAlex year 2025 book chapter),abstract,year_implausible
	132	+7123365826,Event-Triggered Zero-Sum Game for Input Saturated Nonlinear Systems With State Constraints,2026,article,reinforcement_learning,10.1109/tce.2026.3652706,,55.0,155.3,theory,Event-triggered zero-sum game control for saturated nonlinear systems (control theory),abstract,off_scope
	133	+2991046523,Multi-Agent Reinforcement Learning: A Selective Overview of Theories and Algorithms,2021,book-chapter,reinforcement_learning,10.1007/978-3-030-60990-0_12,1911.10635,530.0,178.4,survey,Selective overview of multi-agent RL theory and algorithms,metadata_only,no_abstract
	134	+2973229164,Continuous control for robot based on deep reinforcement learning,2019,dissertation,reinforcement_learning,10.32657/10356/90191,,47.0,171.2,method,Dissertation on deep RL for continuous robot control,abstract,ok
	135	+2982316857,Grandmaster level in StarCraft II using multi-agent reinforcement learning,2019,article,reinforcement_learning,10.1038/s41586-019-1724-z,,1021.0,155.2,engineering,AlphaStar: grandmaster level in StarCraft II with multi-agent RL,metadata_only,no_abstract
	136	+2902907165,"A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play",2018,article,reinforcement_learning,10.1126/science.aar6404,,733.0,138.3,method,"AlphaZero: one self-play RL algorithm for chess, shogi and Go",abstract,ok
	137	+7133481537,Filter-based predefined-time optimal fault-tolerant consensus control for nonlinear multi-agent systems via reinforcement learning,2026,article,reinforcement_learning,10.1016/j.neucom.2026.133224,,35.0,122.9,theory,Predefined-time fault-tolerant consensus control (control theory),metadata_only,off_scope
	138	+7133226216,The Surprising Effectiveness of PPO in Cooperative Multi-Agent Games,2022,conference-paper,reinforcement_learning,10.52202/068431-1787,2103.01955,482.0,95.8,empirical,MAPPO: PPO is surprisingly strong in cooperative multi-agent benchmarks,abstract,ok
	139	+3041807988,Post-Covid-19 Education and Education Technology Solutionism: a Sellers Market,2020,article,reinforcement_learning,10.1007/s42438-020-00164-x,,322.0,93.8,theory,Critical essay on ed-tech solutionism after Covid-19 (not RL),abstract,off_scope
	140	+2781726626,Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor,2018,preprint,reinforcement_learning,10.48550/arxiv.1801.01290,1801.01290,775.0,90.8,method,Soft actor-critic: off-policy maximum-entropy RL; preprint record,abstract,ok
	141	+2963923407,Addressing Function Approximation Error in Actor-Critic Methods,2018,article,reinforcement_learning,,1802.09477,472.0,89.1,method,TD3: clipped double-Q and delayed policy updates against overestimation,abstract,ok
	142	+7122543576,SMSBased EventTriggered Fault Tolerant Control of Nonlinear Systems Using Adaptive Dynamic Programming,2026,article,reinforcement_learning,10.1002/rnc.70380,,25.0,70.6,theory,ADP-based event-triggered fault-tolerant control (control theory),abstract,off_scope
	143	+2899138465,Learning to Learn with Gradients,2018,dissertation,reinforcement_learning,,,20.0,88.6,method,Dissertation on gradient-based meta-learning (MAML lineage),abstract,ok
	144	+4391631327,Axiom-World: A Pre-Registered Study of Two-Stage Post-Training for Rule-Grounded Planning in a Verifiable Toy World,2024,preprint,reinforcement_learning,10.48550/arxiv.2402.03300,2402.03300,99.0,86.4,empirical,Pre-registered toy study of two-stage post-training; record carries arXiv 2402.03300 (DeepSeekMath) and its citations,abstract,identifier_mismatch
	145	+4409147637,Mastering diverse control tasks through world models,2025,article,reinforcement_learning,10.1038/s41586-025-08744-2,,170.0,57.9,method,DreamerV3: world-model RL mastering 150+ tasks with one configuration (Nature),abstract,ok
	146	+2895974005,Exploration and Exploitation in Reinforcement Learning,2018,dissertation,reinforcement_learning,,,18.0,79.8,empirical,Master's thesis on exploration/exploitation parameters,abstract,normalisation_artefact
	147	+7134057781,A fuzzy-TD3 hybrid reinforcement learning framework for robust trajectory tracking of the Mitsubishi RV-2AJ robotic arm,2026,article,reinforcement_learning,10.1038/s41598-026-42615-8,,21.0,23.4,engineering,Fuzzy-TD3 hybrid control for a 5-DOF manipulator (application),abstract,ok
	148	+2754517384,Deep Reinforcement Learning That Matters,2018,conference-paper,reinforcement_learning,10.1609/aaai.v32i1.11694,1709.06560,452.0,70.0,empirical,Deep RL That Matters: reproducibility and variance of deep RL results,abstract,ok
	149	+2761873684,Rainbow: Combining Improvements in Deep Reinforcement Learning,2018,conference-paper,reinforcement_learning,10.1609/aaai.v32i1.11796,,442.0,68.4,method,"Rainbow: combining six DQN extensions, with ablations",abstract,ok
	150	+7133208539,Direct Preference Optimization: Your Language Model is Secretly a Reward Model,2023,conference-paper,reinforcement_learning,10.52202/075280-2338,,293.0,67.6,method,Direct preference optimisation: closed-form RLHF objective without a separate reward model,metadata_only,no_abstract
	151	+7128436456,Synthesis of model predictive control and reinforcement learning: Survey and classification,2026,article,reinforcement_learning,10.1016/j.arcontrol.2026.101045,2502.02133,19.0,21.0,survey,Survey classifying combinations of MPC and RL,abstract,ok
	152	+2962862931,On a Method to Measure Supervised Multiclass Models Interpretability: Application to Degradation Diagnosis (Short Paper),2024,conference-paper,ml_theory,10.4230/oasics.dx.2024.27,,12586.0,3498.4,method,"Interpretability measure for degradation diagnosis (HAL); c36=12,586 cannot belong to this record",abstract,identifier_mismatch
	153	+1698155719,Support Vector Machines,2020,book-chapter,ml_theory,10.1017/9781108564175.026,,1212.0,971.7,survey,Textbook chapter on support vector machines,abstract,ok
	154	+4288283362,Spiral Time: A Geometric Reframing of Temporal Structure and Its Applications in Machine Learning,2026,article,ml_theory,10.5281/zenodo.20632938,,174.0,576.9,method,'Spiral Time' geometric time encoding with a small LSTM ablation; no venue; NCS36 implausible,abstract,normalisation_artefact
	155	+46659105,"Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation",2020,preprint,ml_theory,10.4230/lipics.esem.2026.21,2010.16061,3906.0,608.2,theory,Title = Powers (2011) evaluation-measures paper (arXiv 2010.16061) but the stored abstract is an LLM-as-judge study of program repair,abstract,abstract_mismatch
	156	+3145506661,Introduction to Machine Learning,2019,book-chapter,ml_theory,10.1007/978-3-030-10546-4_1,,1581.0,491.8,survey,Textbook: Introduction to Machine Learning (Springer Briefs),metadata_only,no_abstract
	157	+2291445757,Support Vector Regression,2024,book-chapter,ml_theory,10.1017/9781009006767.012,,769.0,426.8,survey,Textbook chapter on support vector regression,abstract,ok
	158	+2100415632,Efficient learning by implicit exploration in bandit problems with side observations,2026,preprint,ml_theory,10.48550/arxiv.2604.24555,2604.24555,116.0,667.5,theory,Bandits with side observations (2014 paper carrying a 2026 arXiv id and year),abstract,year_implausible
	159	+1568555062,Equivalence and Synthesis of Causal Models,2022,book-chapter,ml_theory,10.1145/3501714.3501732,,576.0,308.7,theory,"Verma & Pearl: equivalence and synthesis of causal models (reprint chapter, 2022)",abstract,ok
	160	+2751318774,Introduction to Machine Learning,2019,book-chapter,ml_theory,10.1142/9789811201967_0001,,978.0,304.2,survey,Textbook: Introduction to Machine Learning (Alpaydin),abstract,ok
	161	+3208688400,Anomaly Detection: A Survey,2021,book-chapter,ml_theory,10.1007/978-981-16-2102-4_36,,833.0,280.5,survey,Anomaly detection survey (book chapter),metadata_only,no_abstract
	162	+3105488314,Risk bounds for statistical learning,2020,dissertation,ml_theory,,,59.0,272.4,theory,Dissertation on risk bounds for statistical learning,metadata_only,no_abstract
	163	+57706852,Stochastic simultaneous optimistic optimization,2026,preprint,ml_theory,10.48550/arxiv.2604.24537,2604.24537,72.0,414.3,theory,StoSOO stochastic optimistic optimisation (2013 paper carrying a 2026 arXiv id and year),abstract,year_implausible
	164	+2946444988,Towards Accurate Model Selection in Deep Unsupervised Domain Adaptation,2026,preprint,ml_theory,10.48550/arxiv.2606.04665,2606.04665,68.0,2646.0,method,Deep Embedded Validation for UDA model selection (2019 paper carrying a 2026 arXiv id; NCS36 implausible),abstract,year_implausible
	165	+2945976633,Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead,2019,article,ml_theory,10.1038/s42256-019-0048-x,1811.10154,1556.0,236.6,theory,Rudin: argument for interpretable models over post-hoc explanation in high-stakes decisions,metadata_only,no_abstract
	166	+4400823820,Random Forest Algorithm Overview,2024,article,ml_theory,10.58496/bjml/2024/007,,782.0,235.4,survey,Overview of random forests (non-indexed journal),abstract,ok
	167	+2999309192,The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation,2020,article,ml_theory,10.1186/s12864-019-6413-7,,1781.0,222.8,theory,Advantages of MCC over F1 and accuracy for imbalanced binary classification,abstract,ok
	168	+2999615587,From local explanations to global understanding with explainable AI for trees,2020,article,ml_theory,10.1038/s42256-019-0138-9,,1767.0,221.0,method,TreeExplainer: exact SHAP values for tree ensembles (NMI),metadata_only,no_abstract
	169	+2439568532,The Mythos of Model Interpretability,2018,article,ml_theory,10.1145/3233231,,1085.0,204.8,theory,Lipton: the mythos of model interpretability,abstract,ok
	170	+3194464139,Penerapan Algoritma K-Nearest Neighbor untuk Penentuan Resiko Kredit Kepemilikan Kendaraan Bermotor,2020,dissertation,ml_theory,,,21.0,197.5,engineering,Indonesian thesis applying kNN to credit risk; NCS36 implausible for c36=21,abstract,normalisation_artefact
	171	+3041133507,A Comprehensive Survey on Transfer Learning,2020,article,ml_theory,10.1109/jproc.2020.3004555,1911.02685,1538.0,192.4,survey,Comprehensive survey of transfer learning (Proc. IEEE),abstract,ok
	172	+2436219157,Optimization Methods for Large-Scale Machine Learning,2018,article,ml_theory,10.1137/16m1080173,1606.04838,956.0,180.4,survey,"Bottou, Curtis, Nocedal: optimisation methods for large-scale ML (SIAM Review)",abstract,ok
	173	+3135028703,"Machine Learning: Algorithms, Real-World Applications and Research Directions",2021,article,ml_theory,10.1007/s42979-021-00592-x,,1518.0,176.9,survey,Overview of ML algorithms and applications (SN Computer Science),metadata_only,no_abstract
	174	+7124426637,A review of machine learning with small and limited data,2026,article,ml_theory,10.1186/s40537-025-01346-9,,46.0,51.2,survey,Review of machine learning with small and limited data,abstract,ok
	175	+2165856466,Spectral bandits for smooth graph functions,2026,preprint,ml_theory,10.48550/arxiv.2604.18420,2604.18420,45.0,259.0,theory,Spectral bandits for smooth graph functions (2014 paper carrying a 2026 arXiv id and year),abstract,year_implausible
	176	+4402843978,Interpretable machine learning,2020,report,ml_theory,10.58248/pn633,,455.0,158.3,survey,UK POSTnote on interpretable ML,abstract,ok
	177	+2964121744,AutoResearch Lite: a reproducible paper-inspired experiment runner (Adam mechanism reproduction and convergence-speed comparison),2026,article,deep_learning_architectures,10.4230/oasics.dx.2026.15,2108.04578,45775.0,50938.5,engineering,"'AutoResearch Lite' Adam reproduction runner; c36=45,775 and arXiv 2108.04578 cannot belong to this record",abstract,identifier_mismatch
	178	+1710476689,Generative Adversarial Networks,2022,book-chapter,deep_learning_architectures,10.1017/9781108891530.013,,2385.0,1018.2,survey,Textbook chapter on GANs (The Science of Deep Learning),abstract,ok
	179	+4322614756,Identifying Resilient Communities in Road Networks: A Path-Based Embedding Approach,2025,conference-paper,deep_learning_architectures,10.4230/lipics.giscience.2025.9,,1542.0,1370.5,method,"Path-based embedding for road-network community detection; c36=1,542 implausible for a LIPIcs paper",abstract,identifier_mismatch
	180	+2963399829,Mixup: Beyond empirical risk minimization,2024,preprint,deep_learning_architectures,10.57702/dcy1c3gw,1710.09412,773.0,674.3,method,mixup: training on convex combinations of examples (ICLR 2018; OpenAlex year 2024),abstract,year_implausible
	181	+4382203079,Are Transformers Effective for Time Series Forecasting?,2023,conference-paper,deep_learning_architectures,10.1609/aaai.v37i9.26317,2205.13504,2278.0,621.0,empirical,Zeng et al.: simple linear models beat Transformers on long-term time-series forecasting benchmarks,abstract,ok
	182	+4406170795,Accurate predictions on small data with a tabular foundation model,2025,article,deep_learning_architectures,10.1038/s41586-024-08328-6,,974.0,331.8,method,TabPFN: tabular foundation model (prior-data fitted network) for small data (Nature),abstract,ok
	183	+146900863,Neural Information Processing Systems,2018,book-chapter,deep_learning_architectures,10.7551/mitpress/11474.003.0014,,999.0,464.0,survey,NeurIPS proceedings volume record (MIT Press),metadata_only,no_abstract
	184	+4396821501,KAN: Kolmogorov-Arnold Networks,2024,preprint,deep_learning_architectures,10.48550/arxiv.2404.19756,2404.19756,531.0,463.2,method,KAN: Kolmogorov-Arnold networks with learnable spline edge functions; preprint,abstract,ok
	185	+2950067852,A Recurrent Latent Variable Model for Sequential Data,2026,preprint,deep_learning_architectures,10.5281/zenodo.19983308,,119.0,4630.6,method,Variational RNN (NeurIPS 2015) recorded as a 2026 preprint; NCS36 implausible,abstract,year_implausible
	186	+4221150249,Graph Neural Networks for Graphs With Heterophily: A Survey,2026,article,deep_learning_architectures,10.1109/tkde.2026.3680353,2202.07082,107.0,119.1,survey,Survey of GNNs for heterophilic graphs,abstract,ok
	187	+7148890333,Spatial-channel collaborative multi-scale graph interaction deep transfer learning for unsupervised rotating machinery fault diagnosis,2026,article,deep_learning_architectures,10.1016/j.engappai.2026.114691,,102.0,113.0,method,Multi-scale graph transfer learning for rotating-machinery diagnosis,metadata_only,no_abstract
	188	+3177318507,Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting,2021,conference-paper,deep_learning_architectures,10.1609/aaai.v35i12.17325,2012.07436,1344.0,315.3,method,Informer: ProbSparse attention for long-sequence forecasting,abstract,ok
	189	+4389326242,Mamba: Linear-Time Sequence Modeling with Selective State Spaces,2023,preprint,deep_learning_architectures,10.48550/arxiv.2312.00752,2312.00752,912.0,303.2,method,Mamba: selective state-space model with linear-time sequence modelling; preprint,abstract,ok
	190	+7118302872,A frame work to predict the influencing nodes in social networking platform using graph neural network model,2026,article,deep_learning_architectures,10.1007/s13278-025-01574-0,,85.0,87.6,method,GNN + ant colony optimisation for influence maximisation,abstract,ok
	191	+3137147200,A Survey of Quantization Methods for Efficient Neural Network Inference,2022,book-chapter,deep_learning_architectures,10.1201/9781003162810-13,,698.0,298.0,survey,Survey of quantisation methods for efficient inference (book chapter),abstract,ok
	192	+3140854437,"Review of deep learning: concepts, CNN architectures, challenges, applications, future directions",2021,article,deep_learning_architectures,10.1186/s40537-021-00444-8,,2452.0,293.1,survey,Review of deep learning concepts and CNN architectures,abstract,ok
	193	+4210257598,A Comprehensive Survey on Graph Neural Networks,2020,article,deep_learning_architectures,,,2239.0,280.0,survey,Comprehensive survey on GNNs (repository duplicate of the TNNLS paper),abstract,ok
	194	+2907492528,A Comprehensive Survey on Graph Neural Networks,2020,article,deep_learning_architectures,10.1109/tnnls.2020.2978386,1901.00596,2230.0,278.9,survey,Comprehensive survey on GNNs (TNNLS),abstract,ok
	195	+3152893301,Graph neural networks: A review of methods and applications,2020,article,deep_learning_architectures,10.1016/j.aiopen.2021.01.001,1812.08434,2023.0,253.0,survey,GNNs: review of methods and applications (AI Open),abstract,ok
	196	+7110333790,Enriching Location Representation with Detailed Semantic Information,2025,article,deep_learning_architectures,10.4230/lipics.giscience.2025.3,,496.0,169.0,method,CaLLiPer+: POI-name-enriched location embeddings; c36=496 implausible for a LIPIcs paper,abstract,identifier_mismatch
	197	+4385763767,Transformers in Time Series: A Survey,2023,conference-paper,deep_learning_architectures,10.24963/ijcai.2023/759,2202.07125,921.0,251.1,survey,Survey of Transformers in time series,abstract,ok
	198	+3124006607,SIMPLE SPECTRAL GRAPH CONVOLUTION,2026,conference-paper,deep_learning_architectures,10.26190/unsworks/32608,,69.0,529.6,method,Simple spectral graph convolution (ICLR 2021; OpenAlex year 2026),abstract,year_implausible
	199	+4285723986,Graph Attention Networks,2020,book-chapter,deep_learning_architectures,10.1007/978-3-031-01587-8_7,,471.0,244.4,survey,Synthesis lecture on graph attention networks,metadata_only,no_abstract
	200	+2892035503,Deep learning for time series classification: a review,2019,article,deep_learning_architectures,10.1007/s10618-019-00619-1,1809.04356,908.0,192.7,survey,Review of deep learning for time-series classification,metadata_only,no_abstract
	201	+2962776659,Fine-grained Analysis of Sentence Embeddings Using Auxiliary Prediction Tasks,2024,preprint,deep_learning_architectures,10.57702/tcr06vve,,212.0,184.9,empirical,Auxiliary-task analysis of sentence embeddings (ICLR 2017; OpenAlex year 2024),abstract,year_implausible
	202	+4406063062,Responsible artificial intelligence governance: A review and research framework,2025,article,ai_safety,10.1016/j.jsis.2024.101885,,534.0,74.5,survey,Review and research framework for responsible AI governance (IS),abstract,ok
	203	+7130596601,"Human in the loop artificial intelligence in healthcare: applications, outcomes, and implementation challenges",2026,article,ai_safety,10.1016/j.ijmedinf.2026.106362,,65.0,29.6,survey,Review of human-in-the-loop AI in healthcare,abstract,ok
	204	+3084163594,Age of Surveillance Capitalism The Fight for a Human Future at the New Frontier of Power.,2020,article,ai_safety,10.26522/brocked.v29i2.849,,775.0,225.7,survey,Book review of Zuboff's Surveillance Capitalism for educators,abstract,off_scope
	205	+2963808661,Scaling of End-To-End Governance Risk Assessments for AI Systems (Practitioner Track),2025,conference-paper,ai_safety,10.4230/oasics.saia.2024.4,,390.0,126.3,engineering,Governance risk-assessment tooling (OASIcs practitioner track); c36=390 implausible,abstract,identifier_mismatch
	206	+7133236347,"TRiSM for Agentic AI: A review of Trust, Risk, and Security Management in LLM-based Agentic Multi-Agent Systems",2026,article,ai_safety,10.1016/j.aiopen.2026.02.006,2506.04133,51.0,44.5,survey,"TRiSM (trust, risk, security management) for LLM-based multi-agent systems",abstract,ok
	207	+7128782475,Large language models provide unsafe answers to patient-posed medical questions,2026,article,ai_safety,10.1038/s41746-026-02428-5,,51.0,23.2,empirical,Physician-led red-teaming of four chatbots on 222 patient questions (HealthAdvice),abstract,ok
	208	+4406571986,Generative artificial intelligence in higher education: Evidence from an analysis of institutional policies and guidelines,2025,article,ai_safety,10.1016/j.chbah.2025.100121,2402.01659,347.0,301.3,empirical,Analysis of generative-AI policies at 116 US R1 universities,abstract,ok
	209	+2981731882,"Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI",2019,article,ai_safety,10.1016/j.inffus.2019.12.012,1910.10045,1097.0,166.8,survey,"Arrieta et al.: XAI concepts, taxonomies and responsible AI",metadata_only,no_abstract
	210	+4402698709,"The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence",2026,article,ai_safety,10.1016/j.patter.2026.101517,2408.12622,44.0,38.4,survey,"AI Risk Repository: 1,725 risks from 74 taxonomies with causal and domain taxonomies",abstract,ok
	211	+4408314709,Bias recognition and mitigation strategies in artificial intelligence healthcare applications,2025,article,ai_safety,10.1038/s41746-025-01503-7,,303.0,42.3,survey,Review of bias recognition and mitigation in healthcare AI,abstract,ok
	212	+4407243994,"AI Ethics: Integrating Transparency, Fairness, and Privacy in AI Development",2025,article,ai_safety,10.1080/08839514.2025.2463722,,300.0,41.8,theory,Comparative ethical framework (EU/US/China policies),abstract,ok
	213	+4330336443,GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models,2023,preprint,ai_safety,10.48550/arxiv.2303.10130,2303.10130,495.0,146.6,empirical,Title = 'GPTs are GPTs' (arXiv 2303.10130) but the stored abstract describes a TRIPS/Anthropic Economic Index comparison,abstract,abstract_mismatch
	214	+7128600296,Advancing healthcare AI governance through a comprehensive maturity model based on systematic review,2026,review,ai_safety,10.1038/s41746-026-02418-7,,41.0,12.8,survey,Systematic review of 35 healthcare-AI governance frameworks and a maturity model,abstract,ok
	215	+4399450035,Power Hungry Processing: Watts Driving the Cost of AI Deployment?,2024,conference-paper,ai_safety,10.1145/3630106.3658542,2311.16863,359.0,138.5,empirical,Power Hungry Processing: measured inference energy of task-specific vs general-purpose models,abstract,ok
	216	+4318586239,Artificial Intelligence: Definition and Background,2023,book-chapter,ai_safety,10.1007/978-3-031-21448-6_2,,247.0,136.6,survey,Policy book chapter defining AI,abstract,ok
	217	+4360615722,Chatbots in Education and Research: A Critical Examination of Ethical Implications and Solutions,2023,article,ai_safety,10.3390/su15075614,,557.0,130.5,theory,Qualitative examination of ethical implications of chatbots in education,abstract,ok
	218	+2670253439,Explanation in artificial intelligence: Insights from the social sciences,2018,article,ai_safety,10.1016/j.artint.2018.07.007,1706.07269,676.0,127.6,survey,Miller: explanation in AI through the lens of the social sciences,metadata_only,no_abstract
	219	+7118631510,Towards trustworthy AI in Industry 5.0: Ante-hoc interpretability with deep learning,2026,article,ai_safety,10.1016/j.cie.2026.111805,,35.0,38.9,method,Ante-hoc interpretability for Industry 5.0,metadata_only,no_abstract
	220	+7127143535,"Industrial information integration through autonomous AI agents: paradoxical effects on transparency, dehumanization, and responsible retail",2026,article,ai_safety,10.1016/j.jii.2026.101089,,33.0,50.0,empirical,Autonomous AI agents in industrial information integration (paradoxical effects),metadata_only,no_abstract
	221	+4409249263,Ethical challenges and evolving strategies in the integration of artificial intelligence into clinical practice,2025,article,ai_safety,10.1371/journal.pdig.0000810,2412.03576,218.0,14.5,survey,Ethical challenges of AI in clinical practice,abstract,ok
	222	+4366262984,Explainable Artificial Intelligence (XAI): What we know and what is left to attain Trustworthy Artificial Intelligence,2023,article,ai_safety,10.1016/j.inffus.2023.101805,,1197.0,109.9,survey,XAI: what is known and what remains for trustworthy AI (Information Fusion),abstract,ok
	223	+2969896603,A Survey on Bias and Fairness in Machine Learning,2021,article,ai_safety,10.1145/3457607,1908.09635,1794.0,102.3,survey,Survey on bias and fairness in machine learning (CSUR),abstract,ok
	224	+4406152291,Medical large language models are vulnerable to data-poisoning attacks,2025,article,ai_safety,10.1038/s41591-024-03445-1,,199.0,67.8,empirical,Data-poisoning threat assessment: 0.001% of tokens with medical misinformation yields harmful LLMs,abstract,ok
	225	+4410609100,Jailbreaking Black Box Large Language Models in Twenty Queries,2025,conference-paper,ai_safety,10.1109/satml64287.2025.00010,2310.08419,193.0,171.5,method,PAIR: black-box semantic jailbreaks in about twenty queries,abstract,ok
	226	+4411542009,Expressing stigma and inappropriate responses prevents LLMs from safely replacing mental health providers.,2025,conference-paper,ai_safety,10.1145/3715275.3732039,2504.18412,191.0,61.8,empirical,Experiments showing LLMs express stigma and respond inappropriately as replacement therapists,abstract,ok
	227	+4360620450,"Opinion Paper: So what if ChatGPT wrote it? Multidisciplinary perspectives on opportunities, challenges and implications of generative conve",2023,article,llm,10.1016/j.ijinfomgt.2023.102642,,3137.0,113.5,survey,Dwivedi et al.: 43 expert contributions on ChatGPT opportunities and challenges,abstract,ok
	228	+3185341429,"Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing",2022,article,llm,10.1145/3560815,2107.13586,2462.0,253.7,survey,Liu et al.: systematic survey of prompt-based learning,abstract,ok
	229	+4384071683,Large language models encode clinical knowledge,2023,article,llm,10.1038/s41586-023-06291-2,2212.13138,2448.0,38.1,benchmark,Med-PaLM: MultiMedQA benchmark and human evaluation framework for clinical LLM answers,abstract,ok
	230	+4384918448,Behavioral Safety and Context Retention of Large Language Models in a Longitudinal ICU Simulation under Offline Conditions,2023,preprint,llm,10.48550/arxiv.2307.09288,2307.09288,2431.0,88.5,empirical,Longitudinal ICU simulation of 23 open-weight models; record carries arXiv 2307.09288 (Llama 2) and its citations,abstract,identifier_mismatch
	231	+4393065402,A survey on large language model based autonomous agents,2024,article,llm,10.1007/s11704-024-40231-1,2308.11432,1820.0,226.1,survey,Wang et al.: survey of LLM-based autonomous agents,abstract,ok
	232	+4319083882,ChatGPT: five priorities for research,2023,article,llm,10.1038/d41586-023-00288-7,,1590.0,146.0,theory,van Dis et al.: five priorities for research on ChatGPT (Nature comment),metadata_only,no_abstract
	233	+4399790131,The effects of over-reliance on AI dialogue systems on students' cognitive abilities: a systematic review,2024,review,ai_safety,10.1186/s40561-024-00316-7,,1484.0,81.8,survey,Systematic review of over-reliance on AI dialogue systems and cognition,abstract,ok
	234	+3133702157,On the Dangers of Stochastic Parrots,2021,conference-paper,llm,10.1145/3442188.3445922,,1464.0,232.1,theory,Bender et al.: risks of ever-larger language models ('stochastic parrots'),abstract,ok
	235	+4321499901,What if the devil is my guardian angel: ChatGPT as a case study of using chatbots in education,2023,article,llm,10.1186/s40561-023-00237-x,,1389.0,325.5,empirical,Qualitative case study of ChatGPT among early adopters in education,abstract,ok

added literature_synthesis/agent_E/make_bib_part05.py +117 −0

		@@ -0,0 +1,117 @@
	1	+#!/usr/bin/env python3
	2	+"""Agent E, WP44 — build manuscript/references_part05.bib + references_verification_part05.csv.
	3	+Each entry is resolved by a single Crossref (DOI) or arXiv (id) lookup; the identifier comes from the ai_ml landmark CSV
	4	+(column doi / arxiv_id) or from an explicit override when the corpus record has none. Nothing is written without a
	5	+successful lookup. Usage: python3 make_bib_part05.py <repo_root>"""
	6	+import csv, json, sys, re, time, unicodedata, urllib.request, xml.etree.ElementTree as ET
	7	+R = sys.argv[1]
	8	+L = {r["work_id"]: r for r in csv.DictReader(open(f"{R}/literature_synthesis/landmarks/ai_ml_landmarks.csv"))}
	9	+MAIL = "spbou4@icloud.com"
	10	+# work_id: (keyword, override identifier or None)
	11	+W = {
	12	+"7124974472": ("rapidadoption", None), "4323655724": ("chatgptgood", None), "2990704537": ("superglue", "arxiv:1905.00537"),
	13	+"4393138529": ("peftsurvey", "arxiv:2403.14608"), "4410551352": ("aurora", None), "4392947521": ("uni", None), "4402220065": ("chief", None),
	14	+"4401397240": ("biomedgpt", None), "4411932262": ("centaur", None), "7133220561": ("instructblip", "arxiv:2305.06500"),
	15	+"7133250524": ("laion5b", "arxiv:2210.08402"), "2995460200": ("vlbert", "arxiv:1908.08530"), "4393148714": ("t2iadapter", None),
	16	+"4289785095": ("textualinversion", "arxiv:2208.01618"), "7164501397": ("textexploit", None), "4396599905": ("lowqualityfusion", None),
	17	+"4288099666": ("cfg", "arxiv:2207.12598"), "4406472463": ("mattergen", None), "4400770292": ("harmlearning", None),
	18	+"4406728221": ("agenticsurvey", None), "7161734976": ("robin", None), "4412709685": ("virtuallab", None), "4407570917": ("agentsociety", None),
	19	+"4413427262": ("agentsvsagentic", None), "4412203333": ("agentmemory", None), "7130481397": ("clinicalagents", None),
	20	+"2781726626": ("sac", "arxiv:1801.01290"), "2963923407": ("td3", "arxiv:1802.09477"), "2982316857": ("alphastar", None),
	21	+"2761873684": ("rainbow", None), "7133226216": ("mappo", "arxiv:2103.01955"), "4409147637": ("dreamerv3", None),
	22	+"2964083594": ("curiosity", "arxiv:1808.04355"), "2991046523": ("marloverview", None), "2945976633": ("stopexplaining", None),
	23	+"2439568532": ("mythos", None), "2999615587": ("treeexplainer", None), "2999309192": ("mcc", None), "3041133507": ("transfersurvey", None),
	24	+"2436219157": ("optimization", None), "4382203079": ("transformerstsf", None), "3177318507": ("informer", None),
	25	+"4385763767": ("tstransformers", None), "4406170795": ("tabpfn", None), "4396821501": ("kan", "arxiv:2404.19756"),
	26	+"2907492528": ("gnnsurvey", None), "3152893301": ("gnnreview", None), "2963399829": ("mixup", "arxiv:1710.09412"),
	27	+"2981731882": ("xai", None), "2969896603": ("biasfairness", None), "2670253439": ("explanation", None), "4366262984": ("xaitrust", None),
	28	+"4402698709": ("riskrepository", None), "4410609100": ("pair", None), "4406152291": ("poisoning", None), "7128782475": ("unsafeanswers", None),
	29	+"4411542009": ("stigma", None), "4406063062": ("raigovernance", None), "4406571986": ("hepolicies", None), "4366420437": ("rapidreview", None),
	30	+"4384464487": ("students", None), "4383312437": ("policyframework", None), "4389437528": ("chatgptedu", None), "4317910584": ("assessment", None),
	31	+"4322617812": ("ragnarok", None), "4413447271": ("radfm", None), "4415913530": ("titan", None), "4392947532": ("conch", None),
	32	+"4385948838": ("plip", None), "4411099958": ("panderm", None), "7128029160": ("brainiac", None), "7119541039": ("earthfm", None),
	33	+"4384071683": ("clinical", None), "3185341429": ("prompting", None), "4319083882": ("priorities", None), "4390730779": ("agentsdefinitions", None),
	34	+"7133236347": ("trism", None), "4407135816": ("fmradiology", None), "4390690043": ("datacentric", None), "4406472599": ("learningfrommodels", None),
	35	+"4410882873": ("multimodalmedicine", None), "4399790131": ("overreliance", None), "4398243313": ("assessmentreview", None),
	36	+"4409269614": ("integrity", None), "4387355843": ("timellm", "arxiv:2310.01728"), "4393153153": ("timesurl", None), "4402727669": ("mplugowl2", None),
	37	+"7148475777": ("gpt5medical", "arxiv:2508.08224"), "7127319731": ("medr1", None), "4414991289": ("dive", None), "4414023998": ("agenticreview", None),
	38	+"4416209823": ("agenticparadigms", None), "7128165235": ("multiagentai", None), "4408314709": ("biashealthcare", None),
	39	+"7140170686": ("pedagogical", None), "7135409568": ("agentichealthcare", None), "3137147200": ("quantization", None), "4361229539": ("llamaadapter", "arxiv:2303.16199"),
	40	+"7133233472": None, # QLoRA already in references.bib (dettmers2023qlora)
	41	+}
	42	+def get(url):
	43	+ req = urllib.request.Request(url, headers={"User-Agent": f"WP44-agentE (mailto:{MAIL})"})
	44	+ return urllib.request.urlopen(req, timeout=40).read()
	45	+def clean(s):
	46	+ s = unicodedata.normalize("NFKC", s)
	47	+ s = s.replace("‐", "-").replace("‑", "-").replace("‒", "-").replace("–", "--").replace("—", "---").replace("‘", "'").replace("’", "'").replace("“", "``").replace("”", "''").replace(" ", " ").replace("−", "-")
	48	+ s = "".join(c if ord(c) < 0x250 else unicodedata.normalize("NFKD", c)[0] if unicodedata.normalize("NFKD", c)[0].isascii() else "" for c in s)
	49	+ s = s.replace("&", "\\&").replace("%", "\\%").replace("#", "\\#").replace("$", "\\$")
	50	+ return re.sub(r"\s+", " ", s).strip()
	51	+def toks(s): return set(re.findall(r"[a-z0-9]+", unicodedata.normalize("NFKD", s).lower()))
	52	+def crossref(doi):
	53	+ m = json.loads(get(f"https://api.crossref.org/works/{doi}?mailto={MAIL}"))["message"]
	54	+ auth = [a for a in m.get("author", []) if a.get("family")]
	55	+ year = (m.get("issued") or m.get("published") or {}).get("date-parts", [[None]])[0][0]
	56	+ return {"type": m.get("type"), "title": m.get("title", [""])[0], "subtitle": (m.get("subtitle") or [""])[0], "authors": [(a.get("family", ""), a.get("given", "")) for a in auth], "container": (m.get("container-title") or [""])[0],
	57	+ "volume": m.get("volume"), "issue": m.get("issue"), "page": m.get("page"), "year": year, "publisher": m.get("publisher"), "doi": m["DOI"], "event": (m.get("event") or {}).get("name")}
	58	+def arxiv(aid):
	59	+ root = ET.fromstring(get(f"https://export.arxiv.org/api/query?id_list={aid}"))
	60	+ ns = {"a": "http://www.w3.org/2005/Atom"}; e = root.find("a:entry", ns)
	61	+ title = re.sub(r"\s+", " ", e.find("a:title", ns).text.strip())
	62	+ if title.startswith("Error"): raise RuntimeError(aid)
	63	+ authors = [a.find("a:name", ns).text for a in e.findall("a:author", ns)]
	64	+ year = int(e.find("a:published", ns).text[:4])
	65	+ return {"type": "arxiv", "title": title, "authors": [(n.split()[-1], " ".join(n.split()[:-1])) for n in authors], "year": year, "arxiv": aid, "doi": None}
	66	+def bib(key, m):
	67	+ A = " and ".join(f"{clean(f)}, {clean(g)}" if g else clean(f) for f, g in m["authors"]) or "Anonymous"
	68	+ T = clean(m["title"] + (": " + m["subtitle"] if m.get("subtitle") else ""))
	69	+ f = [f" title = {{{{{T}}}}}", f" author = {{{A}}}", f" year = {{{m['year']}}}"]
	70	+ t = m["type"]
	71	+ if t == "arxiv":
	72	+ et = "misc"; f += [f" eprint = {{{m['arxiv']}}}", " archiveprefix = {arXiv}", f" howpublished = {{arXiv preprint arXiv:{m['arxiv']}}}", " note = {Preprint, not peer reviewed}"]
	73	+ elif t == "journal-article":
	74	+ et = "article"; f += [f" journal = {{{clean(m['container'])}}}"]
	75	+ if m.get("volume"): f.append(f" volume = {{{m['volume']}}}")
	76	+ if m.get("issue"): f.append(f" number = {{{m['issue']}}}")
	77	+ if m.get("page"): f.append(f" pages = {{{clean(m['page']).replace('-', '--').replace('----', '--')}}}")
	78	+ elif t == "proceedings-article":
	79	+ et = "inproceedings"; f += [f" booktitle = {{{clean(m['container'] or m.get('event') or 'Proceedings')}}}"]
	80	+ if m.get("page"): f.append(f" pages = {{{clean(m['page']).replace('-', '--').replace('----', '--')}}}")
	81	+ if m.get("publisher"): f.append(f" publisher = {{{clean(m['publisher'])}}}")
	82	+ elif t == "book-chapter":
	83	+ et = "incollection"; f += [f" booktitle = {{{clean(m['container'])}}}"]
	84	+ if m.get("publisher"): f.append(f" publisher = {{{clean(m['publisher'])}}}")
	85	+ if m.get("page"): f.append(f" pages = {{{clean(m['page']).replace('-', '--').replace('----', '--')}}}")
	86	+ elif t == "posted-content":
	87	+ et = "misc"; f += [f" howpublished = {{Preprint ({clean(m.get('publisher') or 'repository')})}}", " note = {Preprint, not peer reviewed}"]
	88	+ elif t in ("book", "monograph", "edited-book", "reference-book"):
	89	+ et = "book"; f += [f" publisher = {{{clean(m.get('publisher') or '')}}}"]
	90	+ elif t == "report":
	91	+ et = "techreport"; f += [f" institution = {{{clean(m.get('publisher') or '')}}}"]
	92	+ else:
	93	+ et = "misc"; f += [f" howpublished = {{{clean(m.get('container') or m.get('publisher') or '')}}}"]
	94	+ if m.get("doi"): f.append(f" doi = {{{m['doi']}}}")
	95	+ return f"@{et}{{{key},\n" + ",\n".join(f) + "\n}\n"
	96	+entries, ver, keys = [], [], {}
	97	+for wid, spec in W.items():
	98	+ if spec is None: continue
	99	+ kw, ovr = spec; r = L[wid]
	100	+ ident = ovr or (f"doi:{r['doi']}" if r["doi"] else (f"arxiv:{r['arxiv_id']}" if r["arxiv_id"] else None))
	101	+ if not ident: print("NO IDENTIFIER", wid, r["title"][:60]); continue
	102	+ try:
	103	+ m = crossref(ident[4:]) if ident.startswith("doi:") else arxiv(ident[6:])
	104	+ except Exception as e:
	105	+ print("LOOKUP FAILED", wid, ident, e); continue
	106	+ surname = re.sub(r"[^a-z]", "", unicodedata.normalize("NFKD", m["authors"][0][0]).encode("ascii", "ignore").decode().lower()) if m["authors"] else "anon"
	107	+ key = f"{surname}{m['year']}{kw}"
	108	+ j = len(toks(r["title"]) & toks(m["title"])) / max(1, len(toks(r["title"]) | toks(m["title"])))
	109	+ match = "yes" if j >= 0.5 else f"check (jaccard={j:.2f})"
	110	+ keys[wid] = key; entries.append(bib(key, m))
	111	+ ver.append({"key": key, "openalex_id": wid, "doi_or_arxiv": ident, "verified_via": "crossref" if ident.startswith("doi:") else "arxiv", "verified_title": clean(m["title"])[:160], "corpus_title": clean(r["title"])[:160], "match": match, "entry_type": m["type"]})
	112	+ print(key, "|", match, "|", clean(m["title"])[:70]); time.sleep(0.3)
	113	+open(f"{R}/manuscript/references_part05.bib", "w").write("% references_part05.bib — Part V (The Artificial Intelligence Transformation), WP44. Generated by\n% literature_synthesis/agent_E/make_bib_part05.py on 2026-11-03: every entry resolved by one Crossref (DOI) or arXiv lookup;\n% see references_verification_part05.csv. Keys of manuscript/references.bib are not redefined here.\n\n" + "\n".join(entries))
	114	+with open(f"{R}/manuscript/references_verification_part05.csv", "w", newline="") as f:
	115	+ w = csv.DictWriter(f, fieldnames=list(ver[0].keys())); w.writeheader(); w.writerows(ver)
	116	+json.dump(keys, open(f"{R}/literature_synthesis/agent_E/bib_keys.json", "w"), indent=1)
	117	+print(len(entries), "entries written")

added literature_synthesis/agent_E/make_contributions.py +264 −0

		@@ -0,0 +1,264 @@
	1	+#!/usr/bin/env python3
	2	+"""Agent E (AI Research Specialist), WP44 Phase 5 — contributions.csv for the ai_ml landmark set.
	3	+Reads literature_synthesis/landmarks/ai_ml_landmarks.csv + ai_ml_abstracts.jsonl; the judgements below were made by the
	4	+specialist agent from the abstract (judged_from=abstract) or from title/venue/type only (metadata_only). data_quality_flag:
	5	+ ok | identifier_mismatch (title/abstract do not match the arXiv id or the citation count attached to the record)
	6	+ | abstract_mismatch (abstract belongs to another work) | year_implausible (OpenAlex year contradicts the known first
	7	+ publication) | normalisation_artefact (tiny c36 but very large NCS36: young cohort / rare document type)
	8	+ | off_scope (classified into an AI category but not an AI contribution) | no_abstract (metadata only).
	9	+Usage: python3 literature_synthesis/agent_E/make_contributions.py <repo_root>"""
	10	+import csv, json, sys, os, unicodedata
	11	+R = sys.argv[1] if len(sys.argv) > 1 else "."
	12	+L = {r["work_id"]: r for r in csv.DictReader(open(f"{R}/literature_synthesis/landmarks/ai_ml_landmarks.csv"))}
	13	+A = {str(json.loads(l)["work_id"]): json.loads(l).get("abstract") or "" for l in open(f"{R}/literature_synthesis/landmarks/ai_ml_abstracts.jsonl")}
	14	+def asc(s): return unicodedata.normalize("NFKD", s).encode("ascii", "ignore").decode()
	15	+# work_id: (contribution_type, description, data_quality_flag)
	16	+J = {
	17	+"7172491899": ("engineering","Text-to-MongoDB query translation (GAT + RAG + T5); record carries arXiv 2005.11401 (RAG, Lewis et al.) and its citations","identifier_mismatch"),
	18	+"7154011589": ("empirical","Portuguese-language essay on LLM pre-training data and copyright; record carries arXiv 2005.14165 (GPT-3) and its citations","identifier_mismatch"),
	19	+"4362515116": ("survey","Survey of LLM techniques (pre-training, adaptation, utilisation, evaluation); journal version of the 2023 arXiv survey","ok"),
	20	+"4288253152": ("engineering","Dissertation: fine-tuned multilingual sentence-transformer for Dewey Decimal classification; NCS36 implausible for a dissertation","normalisation_artefact"),
	21	+"4322718191": ("engineering","LLaMA: 7B-65B open-weight foundation language models trained on public data; the abstract reports LLaMA-13B outperforms GPT-3 on most benchmarks","ok"),
	22	+"4323655724": ("survey","Position paper on opportunities and challenges of LLMs for education (student and teacher perspectives)","ok"),
	23	+"4367369313": ("survey","Practitioner guide to prompt engineering for ChatGPT","ok"),
	24	+"4226278401": ("method","Proposal of a 'companion curriculum' for LM training; record carries arXiv 2203.02155 (InstructGPT) and its citations","identifier_mismatch"),
	25	+"4396531244": ("engineering","Dissertation: multilingual semantic search + LLMs for news retrieval; NCS36 implausible for c36=20","normalisation_artefact"),
	26	+"4221143046": ("theory","Essay on 'theatrical compliance' failure mode; record carries arXiv 2201.11903 (chain-of-thought) and its citations","identifier_mismatch"),
	27	+"4327810158": ("engineering","Title = GPT-4 Technical Report (arXiv 2303.08774) but the stored abstract describes an unrelated 'MFOUR Vibe Framework'","abstract_mismatch"),
	28	+"7135096569": ("empirical","Preregistered field experiment (758 BCG consultants) on the uneven ('jagged') effect of GPT-4 assistance on knowledge-work tasks","ok"),
	29	+"4414281281": ("method","DeepSeek-R1: reasoning incentivised by pure reinforcement learning without human-annotated reasoning traces (Nature version)","ok"),
	30	+"4389437528": ("survey","Review of potential benefits of ChatGPT for teaching and learning","ok"),
	31	+"4292779060": ("theory","Speculative 'Aion Framework' on AI consciousness; c36=13,333 cannot belong to this record","identifier_mismatch"),
	32	+"4360836968": ("empirical","'Sparks of AGI': qualitative experiments on an early GPT-4; contested interpretation; preprint","ok"),
	33	+"4406800520": ("survey","Survey of LLM-based agents (brain/perception/action framework); journal version of arXiv 2309.07864","no_abstract"),
	34	+"4385223296": ("survey","Book chapter on responsible use of ChatGPT in education","no_abstract"),
	35	+"4287689466": ("empirical","LLM essay classification for writing placement (450 essays, four models); c36=722 implausible for an OASIcs paper","identifier_mismatch"),
	36	+"4366420437": ("survey","Rapid review of 50 early articles on ChatGPT in education (Dec 2022-Feb 2023)","ok"),
	37	+"4317910584": ("survey","Commentary on ChatGPT and assessment in higher education","ok"),
	38	+"4313564799": ("empirical","SSRN preprint on ChatGPT user experience in education","no_abstract"),
	39	+"4382618722": ("engineering","Chatlaw: multi-agent legal assistant with role-aligned mixture-of-experts (Chinese law)","ok"),
	40	+"4384464487": ("empirical","Survey of 399 Hong Kong students on generative AI perceptions","ok"),
	41	+"4392773210": ("survey","Discussion of AI literacy, prompt engineering and critical thinking in classrooms","ok"),
	42	+"4387355843": ("method","Time-LLM: reprogramming LLMs for time-series forecasting (book-chapter record of the ICLR 2024 paper)","ok"),
	43	+"4392947521": ("method","UNI: general-purpose self-supervised foundation model for computational pathology","no_abstract"),
	44	+"7124890304": ("method","Semi-supervised cross-domain fault diagnosis via contrastive pre-training","no_abstract"),
	45	+"4410551352": ("method","Aurora: foundation model for the Earth system trained on >1M hours of geophysical data; abstract reports outperforming operational forecasts","ok"),
	46	+"4322832366": ("method","TimeMAE: decoupled masked autoencoders for time-series representation","ok"),
	47	+"7133233472": ("method","QLoRA: 4-bit quantised base + low-rank adapters; 65B model finetuned on one 48 GB GPU","ok"),
	48	+"2990704537": ("benchmark","SuperGLUE: harder successor to GLUE after models exceeded non-expert human level","ok"),
	49	+"7125271461": ("engineering","ChemBERTa-3: open training framework for chemical foundation models","ok"),
	50	+"7119541039": ("theory","Position paper defining eleven desirable features of Earth foundation models","ok"),
	51	+"4393138529": ("survey","Survey of parameter-efficient fine-tuning (PEFT) for large models; preprint","ok"),
	52	+"4390690043": ("survey","Survey of data-centric foundation-model approaches in computational healthcare","ok"),
	53	+"7128029160": ("method","BrainIAC: self-supervised foundation model for brain MRI","ok"),
	54	+"4402220065": ("method","Pathology foundation model for cancer diagnosis and prognosis (Nature 2024)","no_abstract"),
	55	+"7118170538": ("survey","Review of AI applications in fisheries","ok"),
	56	+"4411932262": ("method","Centaur: language model fine-tuned on Psych-101 (10M choices) predicting human behaviour across experiments","ok"),
	57	+"4411099958": ("method","PanDerm: multimodal dermatology foundation model pretrained on 2M images, 28 benchmarks","ok"),
	58	+"4361229539": ("method","LLaMA-Adapter: 1.2M-parameter adaption prompts with zero-init attention for instruction following; preprint","ok"),
	59	+"4407135816": ("survey","Narrative review of foundation models for radiology","ok"),
	60	+"4385948838": ("method","PLIP: visual-language pathology foundation model trained on medical Twitter data","no_abstract"),
	61	+"4406643700": ("survey","Quantitative literature analysis and agenda for geospatial AI","ok"),
	62	+"4406472599": ("survey","Perspective on 'learning from models' beyond fine-tuning","no_abstract"),
	63	+"4393153153": ("method","TimesURL: self-supervised contrastive universal time-series representations","ok"),
	64	+"4402727669": ("method","mPLUG-Owl2: modality-collaborative multimodal LLM with shared modules","ok"),
	65	+"4401397240": ("method","BiomedGPT: generalist vision-language foundation model for biomedical tasks","no_abstract"),
	66	+"4385285835": ("method","Masked image modelling at scale for histopathology; medRxiv preprint","ok"),
	67	+"4318718936": ("method","BLIP-2: Q-Former bridging frozen image encoders and frozen LLMs; preprint","ok"),
	68	+"4393148714": ("method","T2I-Adapter: lightweight adapters adding structural/colour control to text-to-image diffusion","ok"),
	69	+"4396599905": ("survey","Survey of multimodal fusion under low-quality data (noise, missing modalities, imbalance, dynamics)","ok"),
	70	+"4396952261": ("survey","Textbook on multimodal deep learning","no_abstract"),
	71	+"7133250524": ("benchmark","LAION-5B: open 5.85B image-text pair dataset for training CLIP-style models","no_abstract"),
	72	+"4392947532": ("method","CONCH: visual-language foundation model for computational pathology","no_abstract"),
	73	+"7138474371": ("method","OpenDriveVLA: vision-language-action model for end-to-end driving on open LLMs","ok"),
	74	+"7127319731": ("method","Med-R1: RL-enhanced vision-language model for medical reasoning","ok"),
	75	+"2619383789": ("survey","Taxonomy of multimodal machine learning (representation, translation, alignment, fusion, co-learning)","ok"),
	76	+"7148475777": ("empirical","Zero-shot evaluation of GPT-5 on multimodal medical QA/VQA","ok"),
	77	+"7160627408": ("method","Generative cross-domain knowledge synthesis framework (conference paper)","ok"),
	78	+"2995460200": ("method","VL-BERT: single-stream Transformer pre-trained on image regions and words","ok"),
	79	+"4415913530": ("method","TITAN: multimodal whole-slide pathology foundation model (335,645 WSIs)","ok"),
	80	+"4289785095": ("method","Textual inversion: personalising text-to-image diffusion with learned pseudo-words from 3-5 images; preprint","ok"),
	81	+"7125233009": ("method","SemCap: sentiment-aware captions as visual proxy for multimodal aspect-based sentiment analysis","ok"),
	82	+"7143512276": ("survey","Survey of multimodal ML in predictive healthcare analytics","ok"),
	83	+"4413447271": ("method","RadFM: generalist radiology foundation model on 2D/3D data (MedMD, 13M images)","ok"),
	84	+"7131830459": ("survey","Survey of multimodal recommender systems","ok"),
	85	+"7206206818": ("method","Lingshu: generalist medical multimodal LLM","ok"),
	86	+"7133220561": ("method","InstructBLIP: instruction tuning of vision-language models","no_abstract"),
	87	+"7164501397": ("empirical","Evaluation of 8 multimodal foundation models on 1,090 medical cases: image predictions largely driven by text","ok"),
	88	+"7165561001": ("method","Hierarchical graph-Mamba multi-agent clinical architecture","no_abstract"),
	89	+"7138470740": ("method","CoEvo: training-free test-time OOD detection with vision-language models","ok"),
	90	+"7138363754": ("method","AdaptCLIP: adapting CLIP for universal visual anomaly detection","ok"),
	91	+"4410882873": ("survey","Scoping review of 432 multimodal medical AI papers (2018-2024)","ok"),
	92	+"4400228804": ("empirical","Dissertation on generative AI in animation production; c36=2 with NCS36=540 is a normalisation artefact","normalisation_artefact"),
	93	+"7124262017": ("survey","OECD Digital Education Outlook 2026 (policy report on generative AI in education)","ok"),
	94	+"4406472463": ("method","MatterGen: diffusion model generating stable inorganic materials under property constraints","ok"),
	95	+"7123342962": ("method","Text-guided optimal transport for unpaired rain removal","ok"),
	96	+"4383312437": ("empirical","AI Ecological Education Policy Framework from 457 students and 180 staff (Hong Kong)","ok"),
	97	+"7133185348": ("method","Imagen: photorealistic text-to-image diffusion with large frozen text encoders","no_abstract"),
	98	+"3195296536": ("method","Dissertation on GAN-based adversarial defence","no_abstract"),
	99	+"4322617812": ("theory","Paradox-theory essay on generative AI in education","ok"),
	100	+"3096831136": ("survey","CACM review of generative adversarial networks","ok"),
	101	+"4400770292": ("empirical","Field experiment: GPT-4 tutoring access improved practice but harmed unassisted exam performance; SSRN preprint","no_abstract"),
	102	+"7127274868": ("empirical","Survey study of parental AI investment and informal English learning","ok"),
	103	+"7124974472": ("empirical","Nationally representative US surveys of generative-AI adoption (45% of adults 18-64 by late 2024)","ok"),
	104	+"7118187240": ("engineering","PatchTST-based photovoltaic forecasting with GAN features (application)","off_scope"),
	105	+"4406199489": ("survey","Review of generative AI and integrity in higher education","ok"),
	106	+"4398243313": ("survey","Scoping review of 32 studies on generative AI and assessment","ok"),
	107	+"4409269614": ("survey","Systematic review (41 studies) of generative AI and academic integrity","ok"),
	108	+"3193917007": ("survey","Textbook chapter on GANs","ok"),
	109	+"4288099666": ("method","Classifier-free diffusion guidance: joint conditional/unconditional training; preprint","ok"),
	110	+"4394934695": ("empirical","Educator interviews on generative AI in higher education","ok"),
	111	+"4321648656": ("survey","Review of neural architectures for AI-generated art","ok"),
	112	+"4395957952": ("survey","UNESCO guidance for generative AI in education and research","no_abstract"),
	113	+"4394863715": ("empirical","SSRN preprint on generative AI and human creativity","no_abstract"),
	114	+"4407778399": ("empirical","Analysis of generative-AI guidelines at 50 US universities","ok"),
	115	+"7124432017": ("theory","Intelligent-TPACK pedagogy framework","ok"),
	116	+"7140170686": ("empirical","Mixed-methods study of human-GenAI pedagogical partnerships","ok"),
	117	+"4406728221": ("survey","Survey of agentic AI concepts and applications","ok"),
	118	+"7161734976": ("engineering","Robin: multi-agent system automating hypothesis generation and data analysis for experimental biology","ok"),
	119	+"7135409568": ("survey","Scoping review of agentic AI in healthcare","ok"),
	120	+"7134187809": ("theory","Event-triggered optimal consensus control for nonlinear multi-agent systems (control theory)","off_scope"),
	121	+"4390730779": ("survey","Survey of LLM-based agents: definitions, methods, prospects","ok"),
	122	+"4225355126": ("survey","Textbook introduction to multi-agent systems","no_abstract"),
	123	+"4407570917": ("engineering","AgentSociety: large-scale simulation of LLM-driven generative social agents","ok"),
	124	+"4410197190": ("survey","Narrative review of agentic AI in organisations","ok"),
	125	+"4413427262": ("survey","Conceptual taxonomy distinguishing AI agents from agentic AI","ok"),
	126	+"4414991289": ("engineering","DIVE: multi-agent workflow extracting experimental data from figures for hydrogen-storage materials","ok"),
	127	+"7122857583": ("theory","Conceptual agentic-AI framework (literature synthesis)","ok"),
	128	+"4401023556": ("survey","Survey of LLM-based multi-agent systems","ok"),
	129	+"4414023998": ("survey","Review of 143 primary studies on agentic AI systems","ok"),
	130	+"7118083939": ("theory","Conceptual overview of agentic information systems","ok"),
	131	+"4412505619": ("survey","Duplicate record of the AI agents vs agentic AI taxonomy (different venue)","ok"),
	132	+"4416209823": ("survey","PRISMA-based review (90 studies) with symbolic vs neural dual-paradigm framework for agentic AI","ok"),
	133	+"7155930820": ("theory","Prescribed-performance consensus control for multi-agent systems (control theory)","off_scope"),
	134	+"7162257503": ("theory","Predefined-time sliding-mode formation control (control theory)","off_scope"),
	135	+"4412709685": ("engineering","Virtual Lab: LLM agents designing SARS-CoV-2 nanobodies with experimental validation","no_abstract"),
	136	+"7128165235": ("theory","Five-component framework for multi-agent AI in knowledge work","ok"),
	137	+"7118190454": ("survey","Perspective on agentic AI for elderly care","ok"),
	138	+"7130481397": ("benchmark","Benchmarking two agent systems (OpenManus, Manus) on clinical decision tasks","ok"),
	139	+"4390638955": ("survey","Scoping review of AI learning tools in K-12 education","off_scope"),
	140	+"4412203333": ("survey","Survey of memory mechanisms of LLM-based agents","ok"),
	141	+"4399982941": ("theory","Dissertation on value engineering for autonomous agents (normative MAS)","normalisation_artefact"),
	142	+"2911283634": ("survey","Textbook chapter on reinforcement learning","no_abstract"),
	143	+"4389473504": ("method","Dissertation on trust-region policy optimisation","ok"),
	144	+"2964083594": ("empirical","Large-scale study of curiosity-driven learning across 54 environments (ICLR 2019; OpenAlex year 2025 is a repository copy)","year_implausible"),
	145	+"2098774185": ("method","Maximum-entropy inverse reinforcement learning (AAAI 2008; OpenAlex year 2018)","year_implausible"),
	146	+"1591675293": ("method","Bayesian inverse reinforcement learning (IJCAI 2007; OpenAlex year 2025 book chapter)","year_implausible"),
	147	+"7123365826": ("theory","Event-triggered zero-sum game control for saturated nonlinear systems (control theory)","off_scope"),
	148	+"2991046523": ("survey","Selective overview of multi-agent RL theory and algorithms","no_abstract"),
	149	+"2973229164": ("method","Dissertation on deep RL for continuous robot control","ok"),
	150	+"2982316857": ("engineering","AlphaStar: grandmaster level in StarCraft II with multi-agent RL","no_abstract"),
	151	+"2902907165": ("method","AlphaZero: one self-play RL algorithm for chess, shogi and Go","ok"),
	152	+"7133481537": ("theory","Predefined-time fault-tolerant consensus control (control theory)","off_scope"),
	153	+"7133226216": ("empirical","MAPPO: PPO is surprisingly strong in cooperative multi-agent benchmarks","ok"),
	154	+"3041807988": ("theory","Critical essay on ed-tech solutionism after Covid-19 (not RL)","off_scope"),
	155	+"2781726626": ("method","Soft actor-critic: off-policy maximum-entropy RL; preprint record","ok"),
	156	+"2963923407": ("method","TD3: clipped double-Q and delayed policy updates against overestimation","ok"),
	157	+"7122543576": ("theory","ADP-based event-triggered fault-tolerant control (control theory)","off_scope"),
	158	+"2899138465": ("method","Dissertation on gradient-based meta-learning (MAML lineage)","ok"),
	159	+"4391631327": ("empirical","Pre-registered toy study of two-stage post-training; record carries arXiv 2402.03300 (DeepSeekMath) and its citations","identifier_mismatch"),
	160	+"4409147637": ("method","DreamerV3: world-model RL mastering 150+ tasks with one configuration (Nature)","ok"),
	161	+"2895974005": ("empirical","Master's thesis on exploration/exploitation parameters","normalisation_artefact"),
	162	+"7134057781": ("engineering","Fuzzy-TD3 hybrid control for a 5-DOF manipulator (application)","ok"),
	163	+"2754517384": ("empirical","Deep RL That Matters: reproducibility and variance of deep RL results","ok"),
	164	+"2761873684": ("method","Rainbow: combining six DQN extensions, with ablations","ok"),
	165	+"7133208539": ("method","Direct preference optimisation: closed-form RLHF objective without a separate reward model","no_abstract"),
	166	+"7128436456": ("survey","Survey classifying combinations of MPC and RL","ok"),
	167	+"2962862931": ("method","Interpretability measure for degradation diagnosis (HAL); c36=12,586 cannot belong to this record","identifier_mismatch"),
	168	+"1698155719": ("survey","Textbook chapter on support vector machines","ok"),
	169	+"4288283362": ("method","'Spiral Time' geometric time encoding with a small LSTM ablation; no venue; NCS36 implausible","normalisation_artefact"),
	170	+"46659105": ("theory","Title = Powers (2011) evaluation-measures paper (arXiv 2010.16061) but the stored abstract is an LLM-as-judge study of program repair","abstract_mismatch"),
	171	+"3145506661": ("survey","Textbook: Introduction to Machine Learning (Springer Briefs)","no_abstract"),
	172	+"2291445757": ("survey","Textbook chapter on support vector regression","ok"),
	173	+"2100415632": ("theory","Bandits with side observations (2014 paper carrying a 2026 arXiv id and year)","year_implausible"),
	174	+"1568555062": ("theory","Verma & Pearl: equivalence and synthesis of causal models (reprint chapter, 2022)","ok"),
	175	+"2751318774": ("survey","Textbook: Introduction to Machine Learning (Alpaydin)","ok"),
	176	+"3208688400": ("survey","Anomaly detection survey (book chapter)","no_abstract"),
	177	+"3105488314": ("theory","Dissertation on risk bounds for statistical learning","no_abstract"),
	178	+"57706852": ("theory","StoSOO stochastic optimistic optimisation (2013 paper carrying a 2026 arXiv id and year)","year_implausible"),
	179	+"2946444988": ("method","Deep Embedded Validation for UDA model selection (2019 paper carrying a 2026 arXiv id; NCS36 implausible)","year_implausible"),
	180	+"2945976633": ("theory","Rudin: argument for interpretable models over post-hoc explanation in high-stakes decisions","no_abstract"),
	181	+"4400823820": ("survey","Overview of random forests (non-indexed journal)","ok"),
	182	+"2999309192": ("theory","Advantages of MCC over F1 and accuracy for imbalanced binary classification","ok"),
	183	+"2999615587": ("method","TreeExplainer: exact SHAP values for tree ensembles (NMI)","no_abstract"),
	184	+"2439568532": ("theory","Lipton: the mythos of model interpretability","ok"),
	185	+"3194464139": ("engineering","Indonesian thesis applying kNN to credit risk; NCS36 implausible for c36=21","normalisation_artefact"),
	186	+"3041133507": ("survey","Comprehensive survey of transfer learning (Proc. IEEE)","ok"),
	187	+"2436219157": ("survey","Bottou, Curtis, Nocedal: optimisation methods for large-scale ML (SIAM Review)","ok"),
	188	+"3135028703": ("survey","Overview of ML algorithms and applications (SN Computer Science)","no_abstract"),
	189	+"7124426637": ("survey","Review of machine learning with small and limited data","ok"),
	190	+"2165856466": ("theory","Spectral bandits for smooth graph functions (2014 paper carrying a 2026 arXiv id and year)","year_implausible"),
	191	+"4402843978": ("survey","UK POSTnote on interpretable ML","ok"),
	192	+"2964121744": ("engineering","'AutoResearch Lite' Adam reproduction runner; c36=45,775 and arXiv 2108.04578 cannot belong to this record","identifier_mismatch"),
	193	+"1710476689": ("survey","Textbook chapter on GANs (The Science of Deep Learning)","ok"),
	194	+"4322614756": ("method","Path-based embedding for road-network community detection; c36=1,542 implausible for a LIPIcs paper","identifier_mismatch"),
	195	+"2963399829": ("method","mixup: training on convex combinations of examples (ICLR 2018; OpenAlex year 2024)","year_implausible"),
	196	+"4382203079": ("empirical","Zeng et al.: simple linear models beat Transformers on long-term time-series forecasting benchmarks","ok"),
	197	+"4406170795": ("method","TabPFN: tabular foundation model (prior-data fitted network) for small data (Nature)","ok"),
	198	+"146900863": ("survey","NeurIPS proceedings volume record (MIT Press)","no_abstract"),
	199	+"4396821501": ("method","KAN: Kolmogorov-Arnold networks with learnable spline edge functions; preprint","ok"),
	200	+"2950067852": ("method","Variational RNN (NeurIPS 2015) recorded as a 2026 preprint; NCS36 implausible","year_implausible"),
	201	+"4221150249": ("survey","Survey of GNNs for heterophilic graphs","ok"),
	202	+"7148890333": ("method","Multi-scale graph transfer learning for rotating-machinery diagnosis","no_abstract"),
	203	+"3177318507": ("method","Informer: ProbSparse attention for long-sequence forecasting","ok"),
	204	+"4389326242": ("method","Mamba: selective state-space model with linear-time sequence modelling; preprint","ok"),
	205	+"7118302872": ("method","GNN + ant colony optimisation for influence maximisation","ok"),
	206	+"3137147200": ("survey","Survey of quantisation methods for efficient inference (book chapter)","ok"),
	207	+"3140854437": ("survey","Review of deep learning concepts and CNN architectures","ok"),
	208	+"4210257598": ("survey","Comprehensive survey on GNNs (repository duplicate of the TNNLS paper)","ok"),
	209	+"2907492528": ("survey","Comprehensive survey on GNNs (TNNLS)","ok"),
	210	+"3152893301": ("survey","GNNs: review of methods and applications (AI Open)","ok"),
	211	+"7110333790": ("method","CaLLiPer+: POI-name-enriched location embeddings; c36=496 implausible for a LIPIcs paper","identifier_mismatch"),
	212	+"4385763767": ("survey","Survey of Transformers in time series","ok"),
	213	+"3124006607": ("method","Simple spectral graph convolution (ICLR 2021; OpenAlex year 2026)","year_implausible"),
	214	+"4285723986": ("survey","Synthesis lecture on graph attention networks","no_abstract"),
	215	+"2892035503": ("survey","Review of deep learning for time-series classification","no_abstract"),
	216	+"2962776659": ("empirical","Auxiliary-task analysis of sentence embeddings (ICLR 2017; OpenAlex year 2024)","year_implausible"),
	217	+"4406063062": ("survey","Review and research framework for responsible AI governance (IS)","ok"),
	218	+"7130596601": ("survey","Review of human-in-the-loop AI in healthcare","ok"),
	219	+"3084163594": ("survey","Book review of Zuboff's Surveillance Capitalism for educators","off_scope"),
	220	+"2963808661": ("engineering","Governance risk-assessment tooling (OASIcs practitioner track); c36=390 implausible","identifier_mismatch"),
	221	+"7133236347": ("survey","TRiSM (trust, risk, security management) for LLM-based multi-agent systems","ok"),
	222	+"7128782475": ("empirical","Physician-led red-teaming of four chatbots on 222 patient questions (HealthAdvice)","ok"),
	223	+"4406571986": ("empirical","Analysis of generative-AI policies at 116 US R1 universities","ok"),
	224	+"2981731882": ("survey","Arrieta et al.: XAI concepts, taxonomies and responsible AI","no_abstract"),
	225	+"4402698709": ("survey","AI Risk Repository: 1,725 risks from 74 taxonomies with causal and domain taxonomies","ok"),
	226	+"4408314709": ("survey","Review of bias recognition and mitigation in healthcare AI","ok"),
	227	+"4407243994": ("theory","Comparative ethical framework (EU/US/China policies)","ok"),
	228	+"4330336443": ("empirical","Title = 'GPTs are GPTs' (arXiv 2303.10130) but the stored abstract describes a TRIPS/Anthropic Economic Index comparison","abstract_mismatch"),
	229	+"7128600296": ("survey","Systematic review of 35 healthcare-AI governance frameworks and a maturity model","ok"),
	230	+"4399450035": ("empirical","Power Hungry Processing: measured inference energy of task-specific vs general-purpose models","ok"),
	231	+"4318586239": ("survey","Policy book chapter defining AI","ok"),
	232	+"4360615722": ("theory","Qualitative examination of ethical implications of chatbots in education","ok"),
	233	+"2670253439": ("survey","Miller: explanation in AI through the lens of the social sciences","no_abstract"),
	234	+"7118631510": ("method","Ante-hoc interpretability for Industry 5.0","no_abstract"),
	235	+"7127143535": ("empirical","Autonomous AI agents in industrial information integration (paradoxical effects)","no_abstract"),
	236	+"4409249263": ("survey","Ethical challenges of AI in clinical practice","ok"),
	237	+"4366262984": ("survey","XAI: what is known and what remains for trustworthy AI (Information Fusion)","ok"),
	238	+"2969896603": ("survey","Survey on bias and fairness in machine learning (CSUR)","ok"),
	239	+"4406152291": ("empirical","Data-poisoning threat assessment: 0.001% of tokens with medical misinformation yields harmful LLMs","ok"),
	240	+"4410609100": ("method","PAIR: black-box semantic jailbreaks in about twenty queries","ok"),
	241	+"4411542009": ("empirical","Experiments showing LLMs express stigma and respond inappropriately as replacement therapists","ok"),
	242	+"4360620450": ("survey","Dwivedi et al.: 43 expert contributions on ChatGPT opportunities and challenges","ok"),
	243	+"3185341429": ("survey","Liu et al.: systematic survey of prompt-based learning","ok"),
	244	+"4384071683": ("benchmark","Med-PaLM: MultiMedQA benchmark and human evaluation framework for clinical LLM answers","ok"),
	245	+"4384918448": ("empirical","Longitudinal ICU simulation of 23 open-weight models; record carries arXiv 2307.09288 (Llama 2) and its citations","identifier_mismatch"),
	246	+"4393065402": ("survey","Wang et al.: survey of LLM-based autonomous agents","ok"),
	247	+"4319083882": ("theory","van Dis et al.: five priorities for research on ChatGPT (Nature comment)","no_abstract"),
	248	+"4399790131": ("survey","Systematic review of over-reliance on AI dialogue systems and cognition","ok"),
	249	+"3133702157": ("theory","Bender et al.: risks of ever-larger language models ('stochastic parrots')","ok"),
	250	+"4321499901": ("empirical","Qualitative case study of ChatGPT among early adopters in education","ok"),
	251	+}
	252	+rows = []
	253	+for wid, r in L.items():
	254	+ ct, desc, flag = J[wid]
	255	+ has = bool(A.get(wid, "").strip())
	256	+ if flag == "ok" and not has: flag = "no_abstract"
	257	+ rows.append({"openalex_id": wid, "title": asc(r["title"])[:140], "year": r["year"], "type": r["type"], "primary_category": r["primary_category"], "doi": r["doi"], "arxiv_id": r["arxiv_id"], "c36": r["c36"], "ncs36": round(float(r["ncs36"]), 1),
	258	+ "contribution_type": ct, "description": desc, "judged_from": "abstract" if has else "metadata_only", "data_quality_flag": flag})
	259	+missing = set(L) - set(J); assert not missing, missing
	260	+out = f"{R}/literature_synthesis/agent_E/contributions.csv"
	261	+with open(out, "w", newline="") as f:
	262	+ w = csv.DictWriter(f, fieldnames=list(rows[0].keys())); w.writeheader(); w.writerows(rows)
	263	+from collections import Counter
	264	+print(len(rows), "rows ->", out); print(Counter(r["data_quality_flag"] for r in rows)); print(Counter(r["contribution_type"] for r in rows)); print(Counter(r["judged_from"] for r in rows))

added literature_synthesis/agent_H/review.md +111 −0

		@@ -0,0 +1,111 @@
	1	+# Agent H — Security, privacy and cryptography: disciplinary synthesis (Phase 5, 2026-11-03)
	2	+
	3	+Area: `security_privacy` (categories `cybersecurity`, `applied_cryptography`, `blockchain`, `privacy_preserving_computation`,
	4	+`federated_learning`). Inputs: `literature_synthesis/landmarks/security_privacy_landmarks.csv` (125 works, 25 per category) and
	5	+`security_privacy_abstracts.jsonl` (92 abstracts); `results/graph/category_counts_yearly.csv`, `category_flows_yearly.csv`,
	6	+`category_cooccurrence_yearly.csv`; `results/bibliometrics/cagr_by_category.csv`; `results/econometrics/breaks.csv` and
	7	+`its_results.json` (rows `cat_*`). No knowledge-graph extraction file exists for this area (`data/processed/graph/extraction/` is
	8	+empty) and no emerging-topic row concerns these categories, so every statement below rests on metadata, abstracts, or corpus
	9	+statistics; nothing rests on a full text. Numbers are quoted here for the record and enter the manuscript only through the
	10	+`\Sec...` macros appended to `manuscript/tables/numbers.tex` by `tables/make_numbers.py` and the four tables
	11	+`tab_security_counts`, `tab_security_summary`, `tab_security_llm_cooccurrence`, `tab_security_landmarks` (`tables/make_tables.py`).
	12	+
	13	+## 0. Reading rules and what the inputs can and cannot support
	14	+
	15	+The landmark set of this area is an automatic selection (`literature_synthesis/select_landmarks.py`): the top works by
	16	+field-normalised 36-month citations (NCS36) within each primary category, with 2025–2026 works ranked by 12-month citations
	17	+relative to their cohort. Three consequences shape everything that follows. First, the selection is blind to record quality:
	18	+duplicated, misdated and misclassified records rank as high as genuine landmarks when OpenAlex attaches citations to them.
	19	+Second, the NCS36 criterion favours survey and application papers in categories whose citing community is broad (blockchain,
	20	+federated learning), and favours standards and reports in categories where the influential documents are not journal
	21	+articles (applied cryptography). Third, 33 of the 125 records are dated 2026 and ranked on a few months of citations; their
	22	+normalised scores (up to 9,806 for an SSRN preprint without abstract) measure cohort position, not durable influence. The
	23	+synthesis therefore treats the landmark list as a *sample of what is highly cited under the corpus's own metric*, documents
	24	+its artefacts explicitly (Section 7), and complements it with the verified references of `manuscript/references.bib` where
	25	+the automatic selection missed works that any specialist would consider foundational (NIST FIPS 205, prompt-injection and
	26	+jailbreak papers, DP-SGD, secure aggregation). Abstract-level evidence is marked "the abstract reports"; works without
	27	+abstract (33 of 125) are discussed from their metadata only.
	28	+
	29	+## 1. Volume and structural change across the five categories (corpus statistics)
	30	+
	31	+Counts are eligible works carrying the category (multi-label) per publication year, summed over the pre/baseline/post labels
	32	+of `category_counts_yearly.csv`; primary-category counts are reported alongside in `tab_security_counts`. Cybersecurity is the
	33	+largest category of the area (78,082 works in 2022; 139,134 in 2025; 829,538 over 2018–2026), applied cryptography the second
	34	+(29,699 in 2018; 51,759 in 2025), blockchain (10,872 → 33,426), privacy-preserving computation (11,378 → 49,266) and federated
	35	+learning (1,040 → 18,093) follow. Growth profiles differ sharply. Cybersecurity and applied cryptography were nearly flat over
	36	+2018–2022 (multi-label CAGR 3.7 % and 1.9 %) and accelerated after 2023 (24.7 % and 20.4 % over 2023–2025). Privacy-preserving
	37	+computation grew throughout (15.2 % then 39.8 %). Federated learning had the steepest baseline growth of the area (54.3 %) and
	38	+remained the fastest-growing category after 2023 (45.8 %), but its growth rate fell. Blockchain is the exception: its growth was
	39	+front-loaded (16.7 % baseline, 17.3 % post, with the 2020 Bai–Perron break), and its share of primary classifications peaked
	40	+around 2020 (1.76 %) before settling (2.02 % in 2025 on a larger base; `category_share_ci_yearly.csv` places the 2023 share
	41	+above the 2025 one when confidence intervals are read).
	42	+
	43	+The interrupted-time-series estimates at the 2022-12 reference month (`breaks.csv`, HAC p-values) show no level shift for any
	44	+of the five categories except federated learning (level −0.126, p = 0.007), and positive slope changes for cybersecurity
	45	+(+0.015 per month, p < 0.001), applied cryptography (+0.013), privacy-preserving computation (+0.012) and none for blockchain
	46	+(−0.002, p = 0.20) or federated learning (−0.006, p = 0.001, i.e. a deceleration). The placebo distribution over 65 alternative
	47	+event months qualifies these: the slope t-statistic at 2022-12 ranks 3rd of 65 for applied cryptography (placebo p = 0.046),
	48	+4th for cybersecurity (0.062), 9th for privacy-preserving computation (0.138), and is unremarkable for blockchain (0.508) and
	49	+federated learning (0.385); the federated-learning level drop is likewise unremarkable against placebos (0.246). Bai–Perron
	50	+(BIC) dates a single break at 2022-07 for cybersecurity, applied cryptography and privacy-preserving computation — five months
	51	+*before* the reference event — and at 2020-04 (blockchain) and 2020-02 (federated learning). PELT places changes at 2022-11
	52	+and 2024-12 for the two security categories, at 2020-12 and 2023-12 for privacy, and at 2020-02 and 2023-05 for federated
	53	+learning. The donut specification (three months excluded on each side) turns the cybersecurity and applied-cryptography
	54	+level estimates positive and significant (+0.051, p = 0.049; +0.066, p = 0.001) while leaving slopes unchanged, which indicates
	55	+that the months around the reference date are themselves unusual (indexing, not necessarily publication). Alternative event
	56	+dates tell the same story: the o1 date (2024-10) carries a large positive level coefficient for cybersecurity (+0.149, p = 0.001),
	57	+applied cryptography (+0.142) and privacy (+0.198), consistent with a late-2024 acceleration that PELT also detects (2024-12).
	58	+
	59	+The reading that these results support is modest. The two "security" categories and privacy-preserving computation grew faster
	60	+after 2023 than before, the acceleration is dated by the break detectors to mid-2022 and to late 2024 rather than to a single
	61	+month, and the growth is partly compositional (Section 6: the share of their works co-classified with the LLM category rose from
	62	+1.7 % in 2022 to 7.6 % in 2025 for cybersecurity and from 1.4 % to 9.3 % for privacy). Blockchain and federated learning are
	63	+stories of the baseline period: both broke upward in early 2020 and neither shows a post-2022 acceleration. None of this is
	64	+causal evidence about ChatGPT; the manuscript's structural-change design (Part IV) is explicit that the reference month is a
	65	+convention.
	66	+
	67	+## 2. Cybersecurity
	68	+
	69	+**What the landmark set contains.** The 25 cybersecurity landmarks are a heterogeneous set: ten are grey literature (five
	70	+dissertations, four book chapters, one report), five have no DOI, and nine are dated 2026. The highest-ranked record is a 2026
	71	+SSRN preprint on persistent homology as a prompt-injection defence (no abstract; NCS36 9,806 on 252 citations within months),
	72	+which cannot be assessed from metadata and is flagged for PI review. Two records are misclassified: *Fedformer* (a long-term
	73	+time-series forecasting Transformer, reissued as a 2026 CRC chapter) and a 2018 book chapter on *Enigma* (a secure-computation
	74	+platform that belongs to privacy-preserving computation). A 2023 IT Professional record titled "IEEE Symposium on Security and
	75	+Privacy" (853 citations in 36 months, no abstract) is a venue-level record, not a paper. These artefacts are consistent with the
	76	+Part III finding that OpenAlex attaches citations to reprints and container records.
	77	+
	78	+**Pre-window landmarks.** The genuine baseline landmarks are the adversarial-examples chapter of Kurakin, Goodfellow and Bengio
	79	+(a 2018 reprint of the 2017 workshop paper; the abstract reports that adversarial examples survive printing and camera capture,
	80	+i.e. transfer to the physical world), the CIC-IDS2017 intrusion-detection dataset paper (no abstract; 534 citations in 36 months,
	81	+the single most reused benchmark of the network-intrusion literature), two 2018 IEEE surveys on blockchain for IoT and on the
	82	+security and privacy of Bitcoin (both abstracts describe surveys), and a 2018 healthcare-cloud blockchain position paper. The
	83	+baseline cybersecurity literature, as sampled by NCS36, was thus dominated by (i) machine-learning-based detection on shared
	84	+datasets and (ii) blockchain-as-security-solution surveys — not by the microarchitectural attacks (Spectre, Meltdown) that the
	85	+verified reference list records as the field's 2018 landmarks; those are classified in computer architecture or receive their
	86	+citations outside the 36-month window.
	87	+
	88	+**Post-window landmarks and the LLM-security turn.** The post-2023 landmarks with abstracts cluster on large language models
	89	+used *for* security and on the security *of* LLM systems. On the first side: *vul-rag* (ACM TOSEM 2026; the abstract reports
	90	+that LLMs distinguish vulnerable from patched code with 0.06–0.14 accuracy and that knowledge-level retrieval raises accuracy by
	91	+16–24 points), *HuntGPT* (anomaly detection with explainability and an LLM front-end), and a RoBERTa-based phishing detector
	92	+(Computer Networks 2026). On the second side: the ACM Computing Surveys survey of security and privacy challenges of LLMs
	93	+(jailbreaking, data poisoning, PII leakage; `das2025llmsecurity`), and the ACM TOSEM study of the Model Context Protocol
	94	+(a lifecycle of MCP servers in four phases and sixteen activities with a threat taxonomy; `hou2026mcp`). The verified reference
	95	+list adds the works that the automatic selection missed because they are preprints or workshop papers: indirect prompt injection
	96	+(`greshake2023injection`), prompt-injection attacks on integrated applications (`liu2023promptinjection`, preprint), the
	97	+"ignore previous prompt" attack (`perez2022ignore`), the jailbreak analysis of Wei et al. (`wei2023jailbroken`, preprint), the
	98	+universal adversarial suffix (`zou2023universal`, preprint), the one-day-exploit agent study (`fang2024oneday`, preprint),
	99	+PentestGPT (`deng2023pentestgpt`, preprint) and the surveys of Yao et al. and Shayegani et al. The corpus statistics locate this
	100	+turn: works co-classified cybersecurity × LLM rose from 1,337 (2022) to 10,549 (2025) and 14,410 in the incomplete 2026, i.e. from
	101	+1.7 % to 7.6 % and 12.0 % of cybersecurity works; co-classification with AI agents reached 5.3 % in 2026. Citation flows show the
	102	+same thing more weakly: 0.7 % of references from cybersecurity works went to LLM-category works in 2025 (1.1 % in 2026), against
	103	+47.5 % to cybersecurity itself.
	104	+
	105	+**Assessment.** Established: the security-of-ML literature predates the window (adversarial examples, membership inference,
	106	+training-data extraction), and the post-2022 LLM-security literature extends it to a new attack surface (instructions in data,
	107	+tool use, agent protocols). Contested: whether LLM-based detection improves on fine-tuned encoders or classical features at equal
	108	+cost — the vul-rag abstract itself reports very low base accuracy, and the phishing and HuntGPT abstracts report gains on
	109	+specific datasets. Speculative: claims that defences against prompt injection can be made robust; the only landmark on the
	110	+question (the 2026 SSRN preprint) argues a negative result and has not been reviewed. The corpus cannot yet tell whether the
	111	+LLM-security literature is a durable subfield or a labelling wave; its first 36-month citation windows close in 2026–2027.

modified manuscript/figures/fig_category_growth.pdf +0 −0

Binary file not shown.

added manuscript/references_part06.bib +712 −0

		@@ -0,0 +1,712 @@
	1	+% references_part06.bib — WP44 Part VI (Agent F, Systems Specialist). Every entry verified 2026-11-03 by a Crossref or arXiv
	2	+% single lookup (see references_verification_part06.csv). Keys already in references.bib are not redefined here.
	3	+
	4	+@inproceedings{chen2024rcacopilot,
	5	+ title = {Automatic Root Cause Analysis via Large Language Models for Cloud Incidents},
	6	+ author = {Yinfang Chen and Huaibing Xie and Minghua Ma and Yu Kang and Xin Gao and Liu Shi and Yunjie Cao and Xuedong Gao and Hao Fan and Ming Wen and Jun Zeng and Supriyo Ghosh and Xuchao Zhang and Chaoyun Zhang and Qingwei Lin and Saravan Rajmohan and Dongmei Zhang and Tianyin Xu},
	7	+ booktitle = {Proceedings of the Nineteenth European Conference on Computer Systems (EuroSys '24)},
	8	+ year = {2024},
	9	+ pages = {674--688},
	10	+ publisher = {ACM},
	11	+ doi = {10.1145/3627703.3629553},
	12	+ note = {verified: Crossref 2026-11-03}
	13	+}
	14	+
	15	+@article{zhou2024drlcloud,
	16	+ title = {Deep reinforcement learning-based methods for resource scheduling in cloud computing: a review and future directions},
	17	+ author = {Guangyao Zhou and Wenhong Tian and Rajkumar Buyya and Ruini Xue and Liang Song},
	18	+ journal = {Artificial Intelligence Review},
	19	+ volume = {57},
	20	+ number = {5},
	21	+ year = {2024},
	22	+ doi = {10.1007/s10462-024-10756-9},
	23	+ note = {verified: Crossref 2026-11-03}
	24	+}
	25	+
	26	+@article{wang2023road6g,
	27	+ title = {On the Road to {6G}: Visions, Requirements, Key Technologies, and Testbeds},
	28	+ author = {Cheng-Xiang Wang and Xiaohu You and Xiqi Gao and Xiuming Zhu and Zixin Li and Chuan Zhang and Haiming Wang and Yongming Huang and Yunfei Chen and Harald Haas and John S. Thompson and Erik G. Larsson and Marco Di Renzo and Wen Tong and Peiying Zhu and Xuemin Shen and H. Vincent Poor and Lajos Hanzo},
	29	+ journal = {IEEE Communications Surveys \& Tutorials},
	30	+ volume = {25},
	31	+ number = {2},
	32	+ pages = {905--974},
	33	+ year = {2023},
	34	+ doi = {10.1109/COMST.2023.3249835},
	35	+ note = {verified: Crossref 2026-11-03}
	36	+}
	37	+
	38	+@article{liu2022isac,
	39	+ title = {Integrated Sensing and Communications: Toward Dual-Functional Wireless Networks for {6G} and Beyond},
	40	+ author = {Fan Liu and Yuanhao Cui and Christos Masouros and Jie Xu and Tony Xiao Han and Yonina C. Eldar and Stefano Buzzi},
	41	+ journal = {IEEE Journal on Selected Areas in Communications},
	42	+ volume = {40},
	43	+ number = {6},
	44	+ pages = {1728--1767},
	45	+ year = {2022},
	46	+ doi = {10.1109/JSAC.2022.3156632},
	47	+ note = {verified: Crossref 2026-11-03}
	48	+}
	49	+
	50	+@article{qu2025meillm,
	51	+ title = {Mobile Edge Intelligence for Large Language Models: A Contemporary Survey},
	52	+ author = {Guanqiao Qu and Qiyuan Chen and Wei Wei and Zheng Lin and Xianhao Chen and Kaibin Huang},
	53	+ journal = {IEEE Communications Surveys \& Tutorials},
	54	+ volume = {27},
	55	+ number = {6},
	56	+ pages = {3820--3860},
	57	+ year = {2025},
	58	+ doi = {10.1109/COMST.2025.3527641},
	59	+ note = {verified: Crossref 2026-11-03}
	60	+}
	61	+
	62	+@article{zhang2026edgegeneral,
	63	+ title = {Toward Edge General Intelligence With Agentic {AI} and Agentification: Concepts, Technologies, and Future Directions},
	64	+ author = {Ruichen Zhang and Guangyuan Liu and Yinqiu Liu and Changyuan Zhao and Jiacheng Wang and Yunting Xu and Dusit Niyato and Jiawen Kang and Yonghui Li and Shiwen Mao and Sumei Sun and Xuemin Shen and Dong In Kim},
	65	+ journal = {IEEE Communications Surveys \& Tutorials},
	66	+ volume = {28},
	67	+ pages = {4285--4318},
	68	+ year = {2026},
	69	+ doi = {10.1109/COMST.2026.3651702},
	70	+ note = {verified: Crossref 2026-11-03}
	71	+}
	72	+
	73	+@article{liu2026eccc,
	74	+ title = {Edge-Cloud Collaborative Computing on Distributed Intelligence and Model Optimization: A Survey},
	75	+ author = {Jing Liu and Yao Du and Kun Yang and Jiaqi Wu and Yan Wang and Xiping Hu and Zehua Wang and Yang Liu and Peng Sun and Azzedine Boukerche and Victor C. M. Leung},
	76	+ journal = {IEEE Communications Surveys \& Tutorials},
	77	+ volume = {28},
	78	+ pages = {5049--5080},
	79	+ year = {2026},
	80	+ doi = {10.1109/COMST.2026.3669216},
	81	+ note = {verified: Crossref 2026-11-03}
	82	+}
	83	+
	84	+@article{jiang2026lamagentic,
	85	+ title = {From Large {AI} Models to Agentic {AI}: A Tutorial on Future Intelligent Communications},
	86	+ author = {Feibo Jiang and Cunhua Pan and Kezhi Wang and Pietro Michiardi and Octavia A. Dobre and Merouane Debbah},
	87	+ journal = {IEEE Journal on Selected Areas in Communications},
	88	+ volume = {44},
	89	+ pages = {3507--3540},
	90	+ year = {2026},
	91	+ doi = {10.1109/JSAC.2026.3660010},
	92	+ note = {verified: Crossref 2026-11-03}
	93	+}
	94	+
	95	+@article{douze2026faiss,
	96	+ title = {The {Faiss} Library},
	97	+ author = {Matthijs Douze and Alexandr Guzhva and Chengqi Deng and Jeff Johnson and Gergely Szilvasy and Pierre-Emmanuel Mazar{\'e} and Maria Lomeli and Lucas Hosseini and Herv{\'e} J{\'e}gou},
	98	+ journal = {IEEE Transactions on Big Data},
	99	+ volume = {12},
	100	+ number = {2},
	101	+ pages = {346--361},
	102	+ year = {2026},
	103	+ doi = {10.1109/TBDATA.2025.3618474},
	104	+ note = {verified: Crossref 2026-11-03}
	105	+}
	106	+
	107	+@article{zha2025datacentric,
	108	+ title = {Data-centric Artificial Intelligence: A Survey},
	109	+ author = {Daochen Zha and Zaid Pervaiz Bhat and Kwei-Herng Lai and Fan Yang and Zhimeng Jiang and Shaochen Zhong and Xia Hu},
	110	+ journal = {ACM Computing Surveys},
	111	+ volume = {57},
	112	+ number = {5},
	113	+ pages = {1--42},
	114	+ year = {2025},
	115	+ doi = {10.1145/3711118},
	116	+ note = {verified: Crossref 2026-11-03}
	117	+}
	118	+
	119	+@article{mohammed2025dataquality,
	120	+ title = {The effects of data quality on machine learning performance on tabular data},
	121	+ author = {Sedir Mohammed and Lukas Budach and Moritz Feuerpfeil and Nina Ihde and Andrea Nathansen and Nele Noack and Hendrik Patzlaff and Felix Naumann and Hazar Harmouch},
	122	+ journal = {Information Systems},
	123	+ volume = {132},
	124	+ pages = {102549},
	125	+ year = {2025},
	126	+ doi = {10.1016/j.is.2025.102549},
	127	+ note = {verified: Crossref 2026-11-03}
	128	+}
	129	+
	130	+@article{dinh2018blockbench,
	131	+ title = {Untangling Blockchain: A Data Processing View of Blockchain Systems},
	132	+ author = {Tien Tuan Anh Dinh and Rui Liu and Meihui Zhang and Gang Chen and Beng Chin Ooi and Ji Wang},
	133	+ journal = {IEEE Transactions on Knowledge and Data Engineering},
	134	+ volume = {30},
	135	+ number = {7},
	136	+ pages = {1366--1385},
	137	+ year = {2018},
	138	+ doi = {10.1109/TKDE.2017.2781227},
	139	+ note = {verified: Crossref 2026-11-03}
	140	+}
	141	+
	142	+@inproceedings{fan2024ragsurvey,
	143	+ title = {A Survey on {RAG} Meeting {LLMs}: Towards Retrieval-Augmented Large Language Models},
	144	+ author = {Wenqi Fan and Yujuan Ding and Liangbo Ning and Shijie Wang and Hengyun Li and Dawei Yin and Tat-Seng Chua and Qing Li},
	145	+ booktitle = {Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD '24)},
	146	+ pages = {6491--6501},
	147	+ year = {2024},
	148	+ publisher = {ACM},
	149	+ doi = {10.1145/3637528.3671470},
	150	+ note = {verified: Crossref 2026-11-03}
	151	+}
	152	+
	153	+@article{wu2024llmrec,
	154	+ title = {A survey on large language models for recommendation},
	155	+ author = {Likang Wu and Zhi Zheng and Zhaopeng Qiu and Hao Wang and Hongchao Gu and Tingjia Shen and Chuan Qin and Chen Zhu and Hengshu Zhu and Qi Liu and Hui Xiong and Enhong Chen},
	156	+ journal = {World Wide Web},
	157	+ volume = {27},
	158	+ number = {5},
	159	+ year = {2024},
	160	+ doi = {10.1007/s11280-024-01291-2},
	161	+ note = {verified: Crossref 2026-11-03}
	162	+}
	163	+
	164	+@article{zhao2024recsysllm,
	165	+ title = {Recommender Systems in the Era of Large Language Models ({LLMs})},
	166	+ author = {Zihuai Zhao and Wenqi Fan and Jiatong Li and Yunqing Liu and Xiaowei Mei and Yiqi Wang and Zhen Wen and Fei Wang and Xiangyu Zhao and Jiliang Tang and Qing Li},
	167	+ journal = {IEEE Transactions on Knowledge and Data Engineering},
	168	+ volume = {36},
	169	+ number = {11},
	170	+ pages = {6889--6907},
	171	+ year = {2024},
	172	+ doi = {10.1109/TKDE.2024.3392335},
	173	+ note = {verified: Crossref 2026-11-03}
	174	+}
	175	+
	176	+@inproceedings{xiao2024cpack,
	177	+ title = {{C-Pack}: Packed Resources For General {Chinese} Embeddings},
	178	+ author = {Shitao Xiao and Zheng Liu and Peitian Zhang and Niklas Muennighoff and Defu Lian and Jian-Yun Nie},
	179	+ booktitle = {Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2024)},
	180	+ pages = {641--649},
	181	+ year = {2024},
	182	+ publisher = {ACM},
	183	+ doi = {10.1145/3626772.3657878},
	184	+ note = {verified: Crossref 2026-11-03}
	185	+}
	186	+
	187	+@article{zhao2026ragaigc,
	188	+ title = {Retrieval-Augmented Generation for {AI}-Generated Content: A Survey},
	189	+ author = {Penghao Zhao and Hailin Zhang and Qinhan Yu and Zhengren Wang and Yunteng Geng and Fangcheng Fu and Ling Yang and Wentao Zhang and Jie Jiang and Bin Cui},
	190	+ journal = {Data Science and Engineering},
	191	+ volume = {11},
	192	+ number = {1},
	193	+ pages = {1--29},
	194	+ year = {2026},
	195	+ doi = {10.1007/s41019-025-00335-5},
	196	+ note = {verified: Crossref 2026-11-03}
	197	+}
	198	+
	199	+@article{huang2026ragsurvey,
	200	+ title = {A Survey on Retrieval-Augmented Text Generation for Large Language Models},
	201	+ author = {Yizheng Huang and Jimmy Xiangji Huang},
	202	+ journal = {ACM Computing Surveys},
	203	+ volume = {58},
	204	+ number = {12},
	205	+ pages = {1--38},
	206	+ year = {2026},
	207	+ doi = {10.1145/3805774},
	208	+ note = {verified: Crossref 2026-11-03}
	209	+}
	210	+
	211	+@inproceedings{he2020lightgcn,
	212	+ title = {{LightGCN}: Simplifying and Powering Graph Convolution Network for Recommendation},
	213	+ author = {Xiangnan He and Kuan Deng and Xiang Wang and Yan Li and YongDong Zhang and Meng Wang},
	214	+ booktitle = {Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '20)},
	215	+ pages = {639--648},
	216	+ year = {2020},
	217	+ publisher = {ACM},
	218	+ doi = {10.1145/3397271.3401063},
	219	+ note = {verified: Crossref 2026-11-03 (Crossref title field is the short form ``LightGCN'')}
	220	+}
	221	+
	222	+@article{wu2023gnnrec,
	223	+ title = {Graph Neural Networks in Recommender Systems: A Survey},
	224	+ author = {Shiwen Wu and Fei Sun and Wentao Zhang and Xu Xie and Bin Cui},
	225	+ journal = {ACM Computing Surveys},
	226	+ volume = {55},
	227	+ number = {5},
	228	+ pages = {1--37},
	229	+ year = {2023},
	230	+ doi = {10.1145/3535101},
	231	+ note = {verified: Crossref 2026-11-03}
	232	+}
	233	+
	234	+@article{gao2023gnnrecsurvey,
	235	+ title = {A Survey of Graph Neural Networks for Recommender Systems: Challenges, Methods, and Directions},
	236	+ author = {Chen Gao and Yu Zheng and Nian Li and Yinfeng Li and Yingrong Qin and Jinghua Piao and Yuhan Quan and Jianxin Chang and Depeng Jin and Xiangnan He and Yong Li},
	237	+ journal = {ACM Transactions on Recommender Systems},
	238	+ volume = {1},
	239	+ number = {1},
	240	+ pages = {1--51},
	241	+ year = {2023},
	242	+ doi = {10.1145/3568022},
	243	+ note = {verified: Crossref 2026-11-03}
	244	+}
	245	+
	246	+@article{davies2018loihi,
	247	+ title = {{Loihi}: A Neuromorphic Manycore Processor with On-Chip Learning},
	248	+ author = {Mike Davies and Narayan Srinivasa and Tsung-Han Lin and Gautham Chinya and Yongqiang Cao and Sri Harsha Choday and Georgios Dimou and Prasad Joshi and Nabil Imam and Shweta Jain and Yuyun Liao and Chit-Kwan Lin and Andrew Lines and Ruokun Liu and Deepak Mathaikutty and Steven McCoy and Arnab Paul and Jonathan Tse and Guruguhanathan Venkataramanan and Yi-Hsin Weng and Andreas Wild and Yoonseok Yang and Hong Wang},
	249	+ journal = {IEEE Micro},
	250	+ volume = {38},
	251	+ number = {1},
	252	+ pages = {82--99},
	253	+ year = {2018},
	254	+ doi = {10.1109/MM.2018.112130359},
	255	+ note = {verified: Crossref 2026-11-03}
	256	+}
	257	+
	258	+@article{schuman2022neuromorphic,
	259	+ title = {Opportunities for neuromorphic computing algorithms and applications},
	260	+ author = {Catherine D. Schuman and Shruti R. Kulkarni and Maryam Parsa and J. Parker Mitchell and Prasanna Date and Bill Kay},
	261	+ journal = {Nature Computational Science},
	262	+ volume = {2},
	263	+ number = {1},
	264	+ pages = {10--19},
	265	+ year = {2022},
	266	+ doi = {10.1038/s43588-021-00184-y},
	267	+ note = {verified: Crossref 2026-11-03}
	268	+}
	269	+
	270	+@article{xu2024taichi,
	271	+ title = {Large-scale photonic chiplet {Taichi} empowers 160-{TOPS/W} artificial general intelligence},
	272	+ author = {Zhihao Xu and Tiankuang Zhou and Muzhou Ma and ChenChen Deng and Qionghai Dai and Lu Fang},
	273	+ journal = {Science},
	274	+ volume = {384},
	275	+ number = {6692},
	276	+ pages = {202--209},
	277	+ year = {2024},
	278	+ doi = {10.1126/science.adl1203},
	279	+ note = {verified: Crossref 2026-11-03}
	280	+}
	281	+
	282	+@article{ahmed2025photonic,
	283	+ title = {Universal photonic artificial intelligence acceleration},
	284	+ author = {Sufi R. Ahmed and Reza Baghdadi and Mikhail Bernadskiy and Nate Bowman and Ryan Braid and Jim Carr and Chen Chen and Pietro Ciccarella and Matthew Cole and John Cooke and Kishor Desai and Carlos Dorta and Jonathan Elmhurst and Bryce Gardiner and others},
	285	+ journal = {Nature},
	286	+ volume = {640},
	287	+ number = {8058},
	288	+ pages = {368--374},
	289	+ year = {2025},
	290	+ doi = {10.1038/s41586-025-08854-x},
	291	+ note = {verified: Crossref 2026-11-03}
	292	+}
	293	+
	294	+@article{fu2024onn,
	295	+ title = {Optical neural networks: progress and challenges},
	296	+ author = {Tingzhao Fu and Jianfa Zhang and Run Sun and Yuyao Huang and Wei Xu and Sigang Yang and Zhihong Zhu and Hongwei Chen},
	297	+ journal = {Light: Science \& Applications},
	298	+ volume = {13},
	299	+ number = {1},
	300	+ year = {2024},
	301	+ doi = {10.1038/s41377-024-01590-3},
	302	+ note = {verified: Crossref 2026-11-03}
	303	+}
	304	+
	305	+@article{lam2026neuromorphicphotonic,
	306	+ title = {Neuromorphic photonic computing with an electro-optic analog memory},
	307	+ author = {Sean Lam and Ahmed Khaled and Simon Bilodeau and Bicky A. Marquez and Paul R. Prucnal and Lukas Chrostowski and Bhavin J. Shastri and Sudip Shekhar},
	308	+ journal = {Nature Communications},
	309	+ volume = {17},
	310	+ number = {1},
	311	+ year = {2026},
	312	+ doi = {10.1038/s41467-026-69084-x},
	313	+ note = {verified: Crossref 2026-11-03}
	314	+}
	315	+
	316	+@inproceedings{patel2024splitwise,
	317	+ title = {{Splitwise}: Efficient Generative {LLM} Inference Using Phase Splitting},
	318	+ author = {Pratyush Patel and Esha Choukse and Chaojie Zhang and Aashaka Shah and {\'I}{\~n}igo Goiri and Saeed Maleki and Ricardo Bianchini},
	319	+ booktitle = {2024 ACM/IEEE 51st Annual International Symposium on Computer Architecture (ISCA)},
	320	+ pages = {118--132},
	321	+ year = {2024},
	322	+ publisher = {IEEE},
	323	+ doi = {10.1109/ISCA59077.2024.00019},
	324	+ note = {verified: Crossref 2026-11-03}
	325	+}
	326	+
	327	+@article{lin2025awq,
	328	+ title = {{AWQ}: Activation-aware Weight Quantization for On-Device {LLM} Compression and Acceleration},
	329	+ author = {Ji Lin and Jiaming Tang and Haotian Tang and Shang Yang and Guangxuan Xiao and Song Han},
	330	+ journal = {GetMobile: Mobile Computing and Communications},
	331	+ volume = {28},
	332	+ number = {4},
	333	+ pages = {12--17},
	334	+ year = {2025},
	335	+ doi = {10.1145/3714983.3714987},
	336	+ note = {verified: Crossref 2026-11-03}
	337	+}
	338	+
	339	+@article{silvano2025accelerators,
	340	+ title = {A Survey on Deep Learning Hardware Accelerators for Heterogeneous {HPC} Platforms},
	341	+ author = {Cristina Silvano and Daniele Ielmini and Fabrizio Ferrandi and Leandro Fiorin and Serena Curzel and Luca Benini and Francesco Conti and Angelo Garofalo and Cristian Zambelli and Enrico Calore and Sebastiano Schifano and Maurizio Palesi and Giuseppe Ascia and Davide Patti and others},
	342	+ journal = {ACM Computing Surveys},
	343	+ volume = {57},
	344	+ number = {11},
	345	+ pages = {1--39},
	346	+ year = {2025},
	347	+ doi = {10.1145/3729215},
	348	+ note = {verified: Crossref 2026-11-03}
	349	+}
	350	+
	351	+@incollection{mutlu2023pim,
	352	+ title = {A Modern Primer on Processing in Memory},
	353	+ author = {Onur Mutlu and Saugata Ghose and Juan G{\'o}mez-Luna and Rachata Ausavarungnirun},
	354	+ booktitle = {Emerging Computing: From Devices to Systems},
	355	+ series = {Computer Architecture and Design Methodologies},
	356	+ pages = {171--243},
	357	+ year = {2023},
	358	+ publisher = {Springer Nature Singapore},
	359	+ doi = {10.1007/978-981-16-7487-7_7},
	360	+ note = {verified: Crossref 2026-11-03}
	361	+}
	362	+
	363	+@article{bennun2020demystifying,
	364	+ title = {Demystifying Parallel and Distributed Deep Learning: An In-depth Concurrency Analysis},
	365	+ author = {Tal Ben-Nun and Torsten Hoefler},
	366	+ journal = {ACM Computing Surveys},
	367	+ volume = {52},
	368	+ number = {4},
	369	+ pages = {1--43},
	370	+ year = {2020},
	371	+ doi = {10.1145/3320060},
	372	+ note = {verified: Crossref 2026-11-03}
	373	+}
	374	+
	375	+@inproceedings{song2024powerinfer,
	376	+ title = {{PowerInfer}: Fast Large Language Model Serving with a Consumer-grade {GPU}},
	377	+ author = {Yixin Song and Zeyu Mi and Haotong Xie and Haibo Chen},
	378	+ booktitle = {Proceedings of the ACM SIGOPS 30th Symposium on Operating Systems Principles (SOSP '24)},
	379	+ pages = {590--606},
	380	+ year = {2024},
	381	+ publisher = {ACM},
	382	+ doi = {10.1145/3694715.3695964},
	383	+ note = {verified: Crossref 2026-11-03}
	384	+}
	385	+
	386	+@inproceedings{strati2024orion,
	387	+ title = {{Orion}: Interference-aware, Fine-grained {GPU} Sharing for {ML} Applications},
	388	+ author = {Foteini Strati and Xianzhe Ma and Ana Klimovic},
	389	+ booktitle = {Proceedings of the Nineteenth European Conference on Computer Systems (EuroSys '24)},
	390	+ pages = {1075--1092},
	391	+ year = {2024},
	392	+ publisher = {ACM},
	393	+ doi = {10.1145/3627703.3629578},
	394	+ note = {verified: Crossref 2026-11-03}
	395	+}
	396	+
	397	+@inproceedings{prabhu2025vattention,
	398	+ title = {{vAttention}: Dynamic Memory Management for Serving {LLMs} without {PagedAttention}},
	399	+ author = {Ramya Prabhu and Ajay Nayak and Jayashree Mohan and Ramachandran Ramjee and Ashish Panwar},
	400	+ booktitle = {Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS '25)},
	401	+ pages = {1133--1150},
	402	+ year = {2025},
	403	+ publisher = {ACM},
	404	+ doi = {10.1145/3669940.3707256},
	405	+ note = {verified: Crossref 2026-11-03}
	406	+}
	407	+
	408	+@inproceedings{lee2020keystone,
	409	+ title = {{Keystone}: an open framework for architecting trusted execution environments},
	410	+ author = {Dayeol Lee and David Kohlbrenner and Shweta Shinde and Krste Asanovi{\'c} and Dawn Song},
	411	+ booktitle = {Proceedings of the Fifteenth European Conference on Computer Systems (EuroSys '20)},
	412	+ pages = {1--16},
	413	+ year = {2020},
	414	+ publisher = {ACM},
	415	+ doi = {10.1145/3342195.3387532},
	416	+ note = {verified: Crossref 2026-11-03 (Crossref title field is the short form ``Keystone'')}
	417	+}
	418	+
	419	+@inproceedings{jia2021nightcore,
	420	+ title = {{Nightcore}: efficient and scalable serverless computing for latency-sensitive, interactive microservices},
	421	+ author = {Zhipeng Jia and Emmett Witchel},
	422	+ booktitle = {Proceedings of the 26th ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS '21)},
	423	+ pages = {152--166},
	424	+ year = {2021},
	425	+ publisher = {ACM},
	426	+ doi = {10.1145/3445814.3446701},
	427	+ note = {verified: Crossref 2026-11-03}
	428	+}
	429	+
	430	+@article{googlequantum2023suppressing,
	431	+ title = {Suppressing quantum errors by scaling a surface code logical qubit},
	432	+ author = {{Google Quantum AI}},
	433	+ journal = {Nature},
	434	+ volume = {614},
	435	+ number = {7949},
	436	+ pages = {676--681},
	437	+ year = {2023},
	438	+ doi = {10.1038/s41586-022-05434-1},
	439	+ note = {verified: Crossref 2026-11-03}
	440	+}
	441	+
	442	+@article{bluvstein2024logical,
	443	+ title = {Logical quantum processor based on reconfigurable atom arrays},
	444	+ author = {Dolev Bluvstein and Simon J. Evered and Alexandra A. Geim and Sophie H. Li and Hengyun Zhou and Tom Manovitz and Sepehr Ebadi and Madelyn Cain and Marcin Kalinowski and Dominik Hangleiter and J. Pablo {Bonilla Ataides} and Nishad Maskara and Iris Cong and Xun Gao and others},
	445	+ journal = {Nature},
	446	+ volume = {626},
	447	+ number = {7997},
	448	+ pages = {58--65},
	449	+ year = {2024},
	450	+ doi = {10.1038/s41586-023-06927-3},
	451	+ note = {verified: Crossref 2026-11-03}
	452	+}
	453	+
	454	+@article{larocca2025barren,
	455	+ title = {Barren plateaus in variational quantum computing},
	456	+ author = {Mart{\'i}n Larocca and Supanut Thanasilp and Samson Wang and Kunal Sharma and Jacob Biamonte and Patrick J. Coles and Lukasz Cincio and Jarrod R. McClean and Zo{\"e} Holmes and M. Cerezo},
	457	+ journal = {Nature Reviews Physics},
	458	+ volume = {7},
	459	+ number = {4},
	460	+ pages = {174--189},
	461	+ year = {2025},
	462	+ doi = {10.1038/s42254-025-00813-9},
	463	+ note = {verified: Crossref 2026-11-03}
	464	+}
	465	+
	466	+@article{main2025distributedqc,
	467	+ title = {Distributed quantum computing across an optical network link},
	468	+ author = {D. Main and P. Drmota and D. P. Nadlinger and E. M. Ainley and A. Agrawal and B. C. Nichol and R. Srinivas and G. Araneda and D. M. Lucas},
	469	+ journal = {Nature},
	470	+ volume = {638},
	471	+ number = {8050},
	472	+ pages = {383--388},
	473	+ year = {2025},
	474	+ doi = {10.1038/s41586-024-08404-x},
	475	+ note = {verified: Crossref 2026-11-03}
	476	+}
	477	+
	478	+@article{aghaee2025aurora,
	479	+ title = {Scaling and networking a modular photonic quantum computer},
	480	+ author = {H. {Aghaee Rad} and T. Ainsworth and R. N. Alexander and B. Altieri and M. F. Askarani and R. Baby and L. Banchi and B. Q. Baragiola and J. E. Bourassa and R. S. Chadwick and I. Charania and H. Chen and M. J. Collins and P. Contu and others},
	481	+ journal = {Nature},
	482	+ volume = {638},
	483	+ number = {8052},
	484	+ pages = {912--919},
	485	+ year = {2025},
	486	+ doi = {10.1038/s41586-024-08406-9},
	487	+ note = {verified: Crossref 2026-11-03}
	488	+}
	489	+
	490	+@misc{javadiabhari2024qiskit,
	491	+ title = {Quantum computing with {Qiskit}},
	492	+ author = {Ali Javadi-Abhari and Matthew Treinish and Kevin Krsulich and Christopher J. Wood and Jake Lishman and Julien Gacon and Simon Martiel and Paul D. Nation and Lev S. Bishop and Andrew W. Cross and Blake R. Johnson and Jay M. Gambetta},
	493	+ year = {2024},
	494	+ eprint = {2405.08810},
	495	+ archiveprefix = {arXiv},
	496	+ howpublished = {arXiv preprint arXiv:2405.08810},
	497	+ note = {verified: arxiv API 2026-11-03}
	498	+}
	499	+
	500	+@article{abughanem2025ibmquantum,
	501	+ title = {{IBM} quantum computers: evolution, performance, and future directions},
	502	+ author = {Muhammad AbuGhanem},
	503	+ journal = {The Journal of Supercomputing},
	504	+ volume = {81},
	505	+ number = {5},
	506	+ year = {2025},
	507	+ doi = {10.1007/s11227-025-07047-7},
	508	+ note = {verified: Crossref 2026-11-03}
	509	+}
	510	+
	511	+@article{kanno2026qsci,
	512	+ title = {Quantum-selected configuration interaction: Classical diagonalization of {Hamiltonians} in subspaces selected by quantum computers},
	513	+ author = {Keita Kanno and Masaya Kohda and Ryosuke Imai and Sho Koh and Kosuke Mitarai and Wataru Mizukami and Yuya O. Nakagawa},
	514	+ journal = {Physical Review Research},
	515	+ volume = {8},
	516	+ number = {2},
	517	+ year = {2026},
	518	+ doi = {10.1103/dmn4-snfx},
	519	+ note = {verified: Crossref 2026-11-03}
	520	+}
	521	+
	522	+@inproceedings{dodge2022carbonintensity,
	523	+ title = {Measuring the Carbon Intensity of {AI} in Cloud Instances},
	524	+ author = {Jesse Dodge and Taylor Prewitt and Remi {Tachet des Combes} and Erika Odmark and Roy Schwartz and Emma Strubell and Alexandra Sasha Luccioni and Noah A. Smith and Nicole DeCario and Will Buchanan},
	525	+ booktitle = {2022 ACM Conference on Fairness, Accountability, and Transparency (FAccT '22)},
	526	+ pages = {1877--1894},
	527	+ year = {2022},
	528	+ publisher = {ACM},
	529	+ doi = {10.1145/3531146.3533234},
	530	+ note = {verified: Crossref 2026-11-03}
	531	+}
	532	+
	533	+@misc{li2023thirsty,
	534	+ title = {Making {AI} Less ``Thirsty'': Uncovering and Addressing the Secret Water Footprint of {AI} Models},
	535	+ author = {Pengfei Li and Jianyi Yang and Mohammad A. Islam and Shaolei Ren},
	536	+ year = {2023},
	537	+ eprint = {2304.03271},
	538	+ archiveprefix = {arXiv},
	539	+ howpublished = {arXiv preprint arXiv:2304.03271},
	540	+ note = {verified: arxiv API 2026-11-03}
	541	+}
	542	+
	543	+@article{li2025thirsty,
	544	+ title = {Making {AI} Less `Thirsty'},
	545	+ author = {Pengfei Li and Jianyi Yang and Mohammad A. Islam and Shaolei Ren},
	546	+ journal = {Communications of the ACM},
	547	+ volume = {68},
	548	+ number = {7},
	549	+ pages = {54--61},
	550	+ year = {2025},
	551	+ doi = {10.1145/3724499},
	552	+ note = {verified: Crossref 2026-11-03}
	553	+}
	554	+
	555	+@inproceedings{chien2023carbongenai,
	556	+ title = {Reducing the Carbon Impact of Generative {AI} Inference (today and in 2035)},
	557	+ author = {Andrew A. Chien and Liuzixuan Lin and Hai Nguyen and Varsha Rao and Tristan Sharma and Rajini Wijayawardana},
	558	+ booktitle = {Proceedings of the 2nd Workshop on Sustainable Computer Systems (HotCarbon '23)},
	559	+ pages = {1--7},
	560	+ year = {2023},
	561	+ publisher = {ACM},
	562	+ doi = {10.1145/3604930.3605705},
	563	+ note = {verified: Crossref 2026-11-03}
	564	+}
	565	+
	566	+@article{boloncanedo2024greenai,
	567	+ title = {A review of green artificial intelligence: Towards a more sustainable future},
	568	+ author = {Ver{\'o}nica Bol{\'o}n-Canedo and Laura Mor{\'a}n-Fern{\'a}ndez and Brais Cancela and Amparo Alonso-Betanzos},
	569	+ journal = {Neurocomputing},
	570	+ volume = {599},
	571	+ pages = {128096},
	572	+ year = {2024},
	573	+ doi = {10.1016/j.neucom.2024.128096},
	574	+ note = {verified: Crossref 2026-11-03}
	575	+}
	576	+
	577	+@article{devriesgao2026datacenters,
	578	+ title = {The carbon and water footprints of data centers and what this could mean for artificial intelligence},
	579	+ author = {Alex de Vries-Gao},
	580	+ journal = {Patterns},
	581	+ volume = {7},
	582	+ number = {1},
	583	+ pages = {101430},
	584	+ year = {2026},
	585	+ doi = {10.1016/j.patter.2025.101430},
	586	+ note = {verified: Crossref 2026-11-03}
	587	+}
	588	+
	589	+@article{thakur2025greenfl,
	590	+ title = {Green Federated Learning: A New Era of Green Aware {AI}},
	591	+ author = {Dipanwita Thakur and Antonella Guzzo and Giancarlo Fortino and Francesco Piccialli},
	592	+ journal = {ACM Computing Surveys},
	593	+ volume = {57},
	594	+ number = {8},
	595	+ pages = {1--36},
	596	+ year = {2025},
	597	+ doi = {10.1145/3718363},
	598	+ note = {verified: Crossref 2026-11-03}
	599	+}
	600	+
	601	+@inproceedings{luccioni2025jevons,
	602	+ title = {From Efficiency Gains to Rebound Effects: The Problem of {Jevons}' Paradox in {AI}'s Polarized Environmental Debate},
	603	+ author = {Alexandra Sasha Luccioni and Emma Strubell and Kate Crawford},
	604	+ booktitle = {Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT '25)},
	605	+ pages = {76--88},
	606	+ year = {2025},
	607	+ publisher = {ACM},
	608	+ doi = {10.1145/3715275.3732007},
	609	+ note = {verified: Crossref 2026-11-03}
	610	+}
	611	+
	612	+@article{chen2026aielectricity,
	613	+ title = {Electricity demand and grid impacts of {AI} data centers: Challenges and prospects},
	614	+ author = {Xin Chen and Xiaoyang Wang and Ana Colacelli and Matt Lee and Le Xie},
	615	+ journal = {Nexus},
	616	+ volume = {3},
	617	+ number = {3},
	618	+ pages = {100162},
	619	+ year = {2026},
	620	+ doi = {10.1016/j.ynexs.2026.100162},
	621	+ note = {verified: Crossref 2026-11-03}
	622	+}
	623	+
	624	+@article{niu2026tokenpowerbench,
	625	+ title = {{TokenPowerBench}: Benchmarking the Power Consumption of {LLM} Inference},
	626	+ author = {Chenxu Niu and Wei Zhang and Jie Li and Yongjian Zhao and Tongyang Wang and Xi Wang and Yong Chen},
	627	+ journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
	628	+ volume = {40},
	629	+ number = {38},
	630	+ pages = {32582--32590},
	631	+ year = {2026},
	632	+ doi = {10.1609/aaai.v40i38.40535},
	633	+ note = {verified: Crossref 2026-11-03}
	634	+}
	635	+
	636	+@article{zhang2019dlwireless,
	637	+ title = {Deep Learning in Mobile and Wireless Networking: A Survey},
	638	+ author = {Chaoyun Zhang and Paul Patras and Hamed Haddadi},
	639	+ journal = {IEEE Communications Surveys \& Tutorials},
	640	+ volume = {21},
	641	+ number = {3},
	642	+ pages = {2224--2287},
	643	+ year = {2019},
	644	+ doi = {10.1109/COMST.2019.2904897},
	645	+ note = {verified: Crossref 2026-11-03}
	646	+}
	647	+
	648	+@techreport{ietf2024rfc9669,
	649	+ title = {{BPF} Instruction Set Architecture ({ISA})},
	650	+ author = {{Internet Engineering Task Force}},
	651	+ institution = {RFC Editor},
	652	+ number = {RFC 9669},
	653	+ year = {2024},
	654	+ doi = {10.17487/RFC9669},
	655	+ note = {verified: Crossref 2026-11-03 (Crossref record carries no author field)}
	656	+}
	657	+
	658	+@inproceedings{liao2026caps,
	659	+ title = {Multi-Objective Scheduling for Large Language Model Inference with Prompt-Level Cost Prediction and {SLO} Awareness},
	660	+ author = {Jiajing Liao and Feng Chang and Yihan Xue and Tianjian Xia and Zeyu Huang and Yuxiao Wang},
	661	+ booktitle = {2026 IEEE 8th International Conference on Communications, Information System and Computer Engineering (CISCE)},
	662	+ pages = {1606--1612},
	663	+ year = {2026},
	664	+ publisher = {IEEE},
	665	+ doi = {10.1109/CISCE69494.2026.11504731},
	666	+ note = {verified: Crossref 2026-11-03}
	667	+}
	668	+
	669	+@inproceedings{li2026deepservecb,
	670	+ title = {{DeepServeCB}: {SLO}-Aware and Cost-Aware Elastic Scheduling for Serverless Multi-Tenant {LLM} Inference},
	671	+ author = {Xuexian Li and Xiayuan Liu and Zilong Wang and Zijian Xiao and Chun-Yao Hsieh and Yixue Liu},
	672	+ booktitle = {2026 International Conference on Generative Artificial Intelligence and Information Security (GAIIS)},
	673	+ pages = {90--95},
	674	+ year = {2026},
	675	+ publisher = {IEEE},
	676	+ doi = {10.1109/GAIIS69281.2026.11519206},
	677	+ note = {verified: Crossref 2026-11-03}
	678	+}
	679	+
	680	+@article{an2026aiflow,
	681	+ title = {{AI} Flow: perspectives, scenarios, and approaches},
	682	+ author = {Hongjun An and Wenhan Hu and Sida Huang and Siqi Huang and Ruanjun Li and Yuanzhi Liang and Jiawei Shao and Yiliang Song and Zihan Wang and Cheng Yuan and Chi Zhang and Hongyuan Zhang and Wenhao Zhuang and Xuelong Li},
	683	+ journal = {Vicinagearth},
	684	+ volume = {3},
	685	+ number = {1},
	686	+ year = {2026},
	687	+ doi = {10.1007/s44336-025-00031-y},
	688	+ note = {verified: Crossref 2026-11-03}
	689	+}
	690	+
	691	+@article{cheikh2026massiveiot,
	692	+ title = {Energy, Scalability, Data, and Security in Massive {IoT}: Current Landscape and Future Directions},
	693	+ author = {Imane Cheikh and S{\'e}bastien Roy and Essaid Sabir and Rachid Aouami},
	694	+ journal = {IEEE Internet of Things Journal},
	695	+ volume = {13},
	696	+ number = {6},
	697	+ pages = {10127--10163},
	698	+ year = {2026},
	699	+ doi = {10.1109/JIOT.2026.3655796},
	700	+ note = {verified: Crossref 2026-11-03}
	701	+}
	702	+
	703	+@article{kumar2026sustainabledc,
	704	+ title = {Toward sustainable data center operation: A review on existing infrastructures, integrated smart energy management, and emerging opportunities},
	705	+ author = {Amit Kumar and Naran M. Pindoriya},
	706	+ journal = {Renewable and Sustainable Energy Reviews},
	707	+ volume = {230},
	708	+ pages = {116664},
	709	+ year = {2026},
	710	+ doi = {10.1016/j.rser.2025.116664},
	711	+ note = {verified: Crossref 2026-11-03 (title verified on its first clause; the subtitle follows the journal record)}
	712	+}

added manuscript/references_part07b.bib +539 −0

		@@ -0,0 +1,539 @@
	1	+% references_part07b.bib — Part VII-b (security half of Part VII: cybersecurity, applied cryptography, blockchain,
	2	+% privacy-preserving computation, federated learning, LLM security). Agent H, 2026-11-03. Every entry verified by a single
	3	+% Crossref, DataCite or arXiv lookup; see references_verification_part07b.csv. Keys that already exist in references.bib
	4	+% (nist2024fips203, nist2024fips204, das2025llmsecurity, greshake2023injection, ...) are reused there and not redefined here.
	5	+
	6	+@techreport{rescorla2018tls13,
	7	+ author = {Rescorla, Eric},
	8	+ title = {{The Transport Layer Security (TLS) Protocol Version 1.3}},
	9	+ year = {2018},
	10	+ institution = {RFC Editor},
	11	+ doi = {10.17487/rfc8446}
	12	+}
	13	+
	14	+@inproceedings{androulaki2018fabric,
	15	+ author = {Androulaki, Elli and Barger, Artem and Bortnikov, Vita and Cachin, Christian and Christidis, Konstantinos and De Caro, Angelo and Enyeart, David and Ferris, Christopher and Laventman, Gennady and Manevich, Yacov and Muralidharan, Srinivasan and Murthy, Chet and Nguyen, Binh and Sethi, Manish and Singh, Gari and Smith, Keith and Sorniotti, Alessandro and Stathakopoulou, Chrysoula and Vukolić, Marko and Cocco, Sharon Weed and Yellick, Jason},
	16	+ title = {{Hyperledger fabric}},
	17	+ year = {2018},
	18	+ booktitle = {Proceedings of the Thirteenth EuroSys Conference},
	19	+ pages = {1--15},
	20	+ publisher = {ACM},
	21	+ doi = {10.1145/3190508.3190538}
	22	+}
	23	+
	24	+@article{yang2019fml,
	25	+ author = {Yang, Qiang and Liu, Yang and Chen, Tianjian and Tong, Yongxin},
	26	+ title = {{Federated Machine Learning}},
	27	+ year = {2019},
	28	+ journal = {ACM Transactions on Intelligent Systems and Technology},
	29	+ volume = {10},
	30	+ number = {2},
	31	+ pages = {1--19},
	32	+ doi = {10.1145/3298981}
	33	+}
	34	+
	35	+@incollection{zhu2020dlg,
	36	+ author = {Zhu, Ligeng and Han, Song},
	37	+ title = {{Deep Leakage from Gradients}},
	38	+ year = {2020},
	39	+ booktitle = {Lecture Notes in Computer Science},
	40	+ pages = {17--31},
	41	+ publisher = {Springer International Publishing},
	42	+ doi = {10.1007/978-3-030-63076-8_2}
	43	+}
	44	+
	45	+@article{tan2023personalizedfl,
	46	+ author = {Tan, Alysa Ziying and Yu, Han and Cui, Lizhen and Yang, Qiang},
	47	+ title = {{Towards Personalized Federated Learning}},
	48	+ year = {2023},
	49	+ journal = {IEEE Transactions on Neural Networks and Learning Systems},
	50	+ volume = {34},
	51	+ number = {12},
	52	+ pages = {9587--9603},
	53	+ doi = {10.1109/tnnls.2022.3160699}
	54	+}
	55	+
	56	+@article{ye2024heterogeneousfl,
	57	+ author = {Ye, Mang and Fang, Xiuwen and Du, Bo and Yuen, Pong C. and Tao, Dacheng},
	58	+ title = {{Heterogeneous Federated Learning: State-of-the-art and Research Challenges}},
	59	+ year = {2024},
	60	+ journal = {ACM Computing Surveys},
	61	+ volume = {56},
	62	+ number = {3},
	63	+ pages = {1--44},
	64	+ doi = {10.1145/3625558}
	65	+}
	66	+
	67	+@article{zhang2023fedala,
	68	+ author = {Zhang, Jianqing and Hua, Yang and Wang, Hao and Song, Tao and Xue, Zhengui and Ma, Ruhui and Guan, Haibing},
	69	+ title = {{FedALA: Adaptive Local Aggregation for Personalized Federated Learning}},
	70	+ year = {2023},
	71	+ journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
	72	+ volume = {37},
	73	+ number = {9},
	74	+ pages = {11237--11244},
	75	+ doi = {10.1609/aaai.v37i9.26330}
	76	+}
	77	+
	78	+@article{nguyen2025unlearning,
	79	+ author = {Nguyen, Thanh Tam and Huynh, Thanh Trung and Ren, Zhao and Nguyen, Phi Le and Liew, Alan Wee-Chung and Yin, Hongzhi and Nguyen, Quoc Viet Hung},
	80	+ title = {{A Survey of Machine Unlearning}},
	81	+ year = {2025},
	82	+ journal = {ACM Transactions on Intelligent Systems and Technology},
	83	+ volume = {16},
	84	+ number = {5},
	85	+ pages = {1--46},
	86	+ doi = {10.1145/3749987}
	87	+}
	88	+
	89	+@article{hou2026mcp,
	90	+ author = {Hou, Xinyi and Zhao, Yanjie and Wang, Shenao and Wang, Haoyu},
	91	+ title = {{Model Context Protocol (MCP): Landscape, Security Threats, and Future Research Directions}},
	92	+ year = {2026},
	93	+ journal = {ACM Transactions on Software Engineering and Methodology},
	94	+ volume = {35},
	95	+ number = {10},
	96	+ pages = {1--37},
	97	+ doi = {10.1145/3796519}
	98	+}
	99	+
	100	+@incollection{castryck2023sidh,
	101	+ author = {Castryck, Wouter and Decru, Thomas},
	102	+ title = {{An Efficient Key Recovery Attack on SIDH}},
	103	+ year = {2023},
	104	+ booktitle = {Lecture Notes in Computer Science},
	105	+ pages = {423--447},
	106	+ publisher = {Springer Nature Switzerland},
	107	+ doi = {10.1007/978-3-031-30589-4_15}
	108	+}
	109	+
	110	+@techreport{nist2019pqcround1,
	111	+ author = {Alagic, Gorjan and Alperin-Sheriff, Jacob and Apon, Daniel and Cooper, David and Dang, Quynh and Liu, Yi-Kai and Miller, Carl and Moody, Dustin and Peralta, Rene and Perlner, Ray and Robinson, Angela and Smith-Tone, Daniel},
	112	+ title = {{Status report on the first round of the NIST post-quantum cryptography standardization process}},
	113	+ year = {2019},
	114	+ institution = {National Institute of Standards and Technology},
	115	+ doi = {10.6028/nist.ir.8240}
	116	+}
	117	+
	118	+@techreport{nist2022pqcround3,
	119	+ author = {Alagic, Gorjan and Apon, Daniel and Cooper, David and Dang, Quynh and Dang, Thinh and Kelsey, John and Lichtinger, Jacob and Liu, Yi-Kai and Miller, Carl and Moody, Dustin and Peralta, Rene and Perlner, Ray and Robinson, Angela and Smith-Tone, Daniel},
	120	+ title = {{Status report on the third round of the NIST Post-Quantum Cryptography Standardization process}},
	121	+ year = {2022},
	122	+ institution = {National Institute of Standards and Technology (U.S.)},
	123	+ doi = {10.6028/nist.ir.8413}
	124	+}
	125	+
	126	+@article{foley2019bitcoin,
	127	+ author = {Foley, Sean and Karlsen, Jonathan R and Putniņš, Tālis J},
	128	+ title = {{Sex, Drugs, and Bitcoin: How Much Illegal Activity Is Financed through Cryptocurrencies?}},
	129	+ year = {2019},
	130	+ journal = {The Review of Financial Studies},
	131	+ volume = {32},
	132	+ number = {5},
	133	+ pages = {1798--1853},
	134	+ doi = {10.1093/rfs/hhz015}
	135	+}
	136	+
	137	+@article{cong2019smartcontracts,
	138	+ author = {Cong, Lin William and He, Zhiguo},
	139	+ title = {{Blockchain Disruption and Smart Contracts}},
	140	+ year = {2019},
	141	+ journal = {The Review of Financial Studies},
	142	+ volume = {32},
	143	+ number = {5},
	144	+ pages = {1754--1797},
	145	+ doi = {10.1093/rfs/hhz007}
	146	+}
	147	+
	148	+@techreport{yaga2018blockchain,
	149	+ author = {Yaga, Dylan and Mell, Peter and Roby, Nik and Scarfone, Karen},
	150	+ title = {{Blockchain technology overview}},
	151	+ year = {2018},
	152	+ institution = {National Institute of Standards and Technology},
	153	+ doi = {10.6028/nist.ir.8202}
	154	+}
	155	+
	156	+@article{wang2019consensus,
	157	+ author = {Wang, Wenbo and Hoang, Dinh Thai and Hu, Peizhao and Xiong, Zehui and Niyato, Dusit and Wang, Ping and Wen, Yonggang and Kim, Dong In},
	158	+ title = {{A Survey on Consensus Mechanisms and Mining Strategy Management in Blockchain Networks}},
	159	+ year = {2019},
	160	+ journal = {IEEE Access},
	161	+ volume = {7},
	162	+ pages = {22328--22370},
	163	+ doi = {10.1109/access.2019.2896108}
	164	+}
	165	+
	166	+@incollection{kurakin2018physical,
	167	+ author = {Kurakin, Alexey and Goodfellow, Ian J. and Bengio, Samy},
	168	+ title = {{Adversarial Examples in the Physical World}},
	169	+ year = {2018},
	170	+ booktitle = {Artificial Intelligence Safety and Security},
	171	+ pages = {99--112},
	172	+ publisher = {Chapman and Hall/CRC},
	173	+ doi = {10.1201/9781351251389-8}
	174	+}
	175	+
	176	+@article{thapa2022splitfed,
	177	+ author = {Thapa, Chandra and Mahawaga Arachchige, Pathum Chamikara and Camtepe, Seyit and Sun, Lichao},
	178	+ title = {{SplitFed: When Federated Learning Meets Split Learning}},
	179	+ year = {2022},
	180	+ journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
	181	+ volume = {36},
	182	+ number = {8},
	183	+ pages = {8485--8493},
	184	+ doi = {10.1609/aaai.v36i8.20825}
	185	+}
	186	+
	187	+@article{du2026vulrag,
	188	+ author = {Du, Xueying and Zheng, Geng and Wang, Kaixin and Zou, Yi and Wang, Yujia and Deng, Wentai and Feng, Jiayi and Liu, Mingwei and Chen, Bihuan and Peng, Xin and Ma, Tao and Lou, Yiling},
	189	+ title = {{vul-rag : Enhancing LLM-based Vulnerability Detection via Knowledge-level RAG}},
	190	+ year = {2026},
	191	+ journal = {ACM Transactions on Software Engineering and Methodology},
	192	+ volume = {35},
	193	+ number = {10},
	194	+ pages = {1--25},
	195	+ doi = {10.1145/3797277}
	196	+}
	197	+
	198	+@article{ahmed2020tracingapps,
	199	+ author = {Ahmed, Nadeem and Michelin, Regio A. and Xue, Wanli and Ruj, Sushmita and Malaney, Robert and Kanhere, Salil S. and Seneviratne, Aruna and Hu, Wen and Janicke, Helge and Jha, Sanjay K.},
	200	+ title = {{A Survey of COVID-19 Contact Tracing Apps}},
	201	+ year = {2020},
	202	+ journal = {IEEE Access},
	203	+ volume = {8},
	204	+ pages = {134577--134601},
	205	+ doi = {10.1109/access.2020.3010226}
	206	+}
	207	+
	208	+@article{sheller2020flmedicine,
	209	+ author = {Sheller, Micah J. and Edwards, Brandon and Reina, G. Anthony and Martin, Jason and Pati, Sarthak and Kotrotsou, Aikaterini and Milchenko, Mikhail and Xu, Weilin and Marcus, Daniel and Colen, Rivka R. and Bakas, Spyridon},
	210	+ title = {{Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data}},
	211	+ year = {2020},
	212	+ journal = {Scientific Reports},
	213	+ volume = {10},
	214	+ number = {1},
	215	+ doi = {10.1038/s41598-020-69250-1}
	216	+}
	217	+
	218	+@article{chen2020fedhealth,
	219	+ author = {Chen, Yiqiang and Qin, Xin and Wang, Jindong and Yu, Chaohui and Gao, Wen},
	220	+ title = {{FedHealth: A Federated Transfer Learning Framework for Wearable Healthcare}},
	221	+ year = {2020},
	222	+ journal = {IEEE Intelligent Systems},
	223	+ volume = {35},
	224	+ number = {4},
	225	+ pages = {83--93},
	226	+ doi = {10.1109/mis.2020.2988604}
	227	+}
	228	+
	229	+@article{zhang2024fedtgp,
	230	+ author = {Zhang, Jianqing and Liu, Yang and Hua, Yang and Cao, Jian},
	231	+ title = {{FedTGP: Trainable Global Prototypes with Adaptive-Margin-Enhanced Contrastive Learning for Data and Model Heterogeneity in Federated Learning}},
	232	+ year = {2024},
	233	+ journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
	234	+ volume = {38},
	235	+ number = {15},
	236	+ pages = {16768--16776},
	237	+ doi = {10.1609/aaai.v38i15.29617}
	238	+}
	239	+
	240	+@article{fu2026dpfl,
	241	+ author = {Fu, Jie and Hong, Yuan and Ling, Xinpeng and Wang, Leixia and Ran, Xun and Sun, Zhiyu and Wang, Wendy Hui and Chen, Zhili and Cao, Yang},
	242	+ title = {{Differentially Private Federated Learning: A Systematic Review}},
	243	+ year = {2026},
	244	+ journal = {ACM Computing Surveys},
	245	+ volume = {58},
	246	+ number = {11},
	247	+ pages = {1--38},
	248	+ doi = {10.1145/3801079}
	249	+}
	250	+
	251	+@article{jimenezgutierrez2026flsecurity,
	252	+ author = {Jimenez-Gutierrez, Daniel M. and Falkouskaya, Yelizaveta and Hernandez-Ramos, José L. and Anagnostopoulos, Aris and Chatzigiannakis, Ioannis and Vitaletti, Andrea},
	253	+ title = {{On the security and privacy of federated learning: A survey with attacks, defenses, frameworks, applications, and future directions}},
	254	+ year = {2026},
	255	+ journal = {Information Fusion},
	256	+ volume = {131},
	257	+ pages = {104155},
	258	+ doi = {10.1016/j.inffus.2026.104155}
	259	+}
	260	+
	261	+@article{davidson2018institutions,
	262	+ author = {Davidson, Sinclair and De Filippi, Primavera and Potts, Jason},
	263	+ title = {{Blockchains and the economic institutions of capitalism}},
	264	+ year = {2018},
	265	+ journal = {Journal of Institutional Economics},
	266	+ volume = {14},
	267	+ number = {4},
	268	+ pages = {639--658},
	269	+ doi = {10.1017/s1744137417000200}
	270	+}
	271	+
	272	+@article{mendling2018bpm,
	273	+ author = {Mendling, Jan and Weber, Ingo and Aalst, Wil Van Der and Brocke, Jan Vom and Cabanillas, Cristina and Daniel, Florian and Debois, Søren and Ciccio, Claudio Di and Dumas, Marlon and Dustdar, Schahram and Gal, Avigdor and García-Bañuelos, Luciano and Governatori, Guido and Hull, Richard and Rosa, Marcello La and Leopold, Henrik and Leymann, Frank and Recker, Jan and Reichert, Manfred and Reijers, Hajo A. and Rinderle-Ma, Stefanie and Solti, Andreas and Rosemann, Michael and Schulte, Stefan and Singh, Munindar P. and Slaats, Tijs and Staples, Mark and Weber, Barbara and Weidlich, Matthias and Weske, Mathias and Xu, Xiwei and Zhu, Liming},
	274	+ title = {{Blockchains for Business Process Management - Challenges and Opportunities}},
	275	+ year = {2018},
	276	+ journal = {ACM Transactions on Management Information Systems},
	277	+ volume = {9},
	278	+ number = {1},
	279	+ pages = {1--16},
	280	+ doi = {10.1145/3183367}
	281	+}
	282	+
	283	+@article{saberi2019supplychain,
	284	+ author = {Saberi, Sara and Kouhizadeh, Mahtab and Sarkis, Joseph and Shen, Lejia},
	285	+ title = {{Blockchain technology and its relationships to sustainable supply chain management}},
	286	+ year = {2019},
	287	+ journal = {International Journal of Production Research},
	288	+ volume = {57},
	289	+ number = {7},
	290	+ pages = {2117--2135},
	291	+ doi = {10.1080/00207543.2018.1533261}
	292	+}
	293	+
	294	+@article{conti2018bitcoinsecurity,
	295	+ author = {Conti, Mauro and Sandeep Kumar, E. and Lal, Chhagan and Ruj, Sushmita},
	296	+ title = {{A Survey on Security and Privacy Issues of Bitcoin}},
	297	+ year = {2018},
	298	+ journal = {IEEE Communications Surveys \& Tutorials},
	299	+ volume = {20},
	300	+ number = {4},
	301	+ pages = {3416--3452},
	302	+ doi = {10.1109/comst.2018.2842460}
	303	+}
	304	+
	305	+@article{fernandez2018biot,
	306	+ author = {Fernandez-Carames, Tiago M. and Fraga-Lamas, Paula},
	307	+ title = {{A Review on the Use of Blockchain for the Internet of Things}},
	308	+ year = {2018},
	309	+ journal = {IEEE Access},
	310	+ volume = {6},
	311	+ pages = {32979--33001},
	312	+ doi = {10.1109/access.2018.2842685}
	313	+}
	314	+
	315	+@article{huynh2023metaverse,
	316	+ author = {Huynh-The, Thien and Gadekallu, Thippa Reddy and Wang, Weizheng and Yenduri, Gokul and Ranaweera, Pasika and Pham, Quoc-Viet and da Costa, Daniel Benevides and Liyanage, Madhusanka},
	317	+ title = {{Blockchain for the metaverse: A Review}},
	318	+ year = {2023},
	319	+ journal = {Future Generation Computer Systems},
	320	+ volume = {143},
	321	+ pages = {401--419},
	322	+ doi = {10.1016/j.future.2023.02.008}
	323	+}
	324	+
	325	+@article{mohammadi2026dpmedical,
	326	+ author = {Mohammadi, Marziyeh and Vejdanihemmat, Mohsen and Lotfinia, Mahshad and Rusu, Mirabela and Truhn, Daniel and Maier, Andreas and Tayebi Arasteh, Soroosh},
	327	+ title = {{Differential privacy for medical deep learning: methods, tradeoffs, and deployment implications}},
	328	+ year = {2026},
	329	+ journal = {npj Digital Medicine},
	330	+ volume = {9},
	331	+ number = {1},
	332	+ doi = {10.1038/s41746-025-02280-z}
	333	+}
	334	+
	335	+@article{eden2025flgovernance,
	336	+ author = {Eden, Rebekah and Chukwudi, Ignatius and Bain, Chris and Barbieri, Sebastiano and Callaway, Leonie and de Jersey, Susan and George, Yasmeen and Gorse, Alain-Dominique and Lawley, Michael and Marendy, Peter and McPhail, Steven M. and Nguyen, Anthony and Samadbeik, Mahnaz and Sullivan, Clair},
	337	+ title = {{A scoping review of the governance of federated learning in healthcare}},
	338	+ year = {2025},
	339	+ journal = {npj Digital Medicine},
	340	+ volume = {8},
	341	+ number = {1},
	342	+ doi = {10.1038/s41746-025-01836-3}
	343	+}
	344	+
	345	+@article{zhu2021noniid,
	346	+ author = {Zhu, Hangyu and Xu, Jinjin and Liu, Shiqing and Jin, Yaochu},
	347	+ title = {{Federated learning on non-IID data: A survey}},
	348	+ year = {2021},
	349	+ journal = {Neurocomputing},
	350	+ volume = {465},
	351	+ pages = {371--390},
	352	+ doi = {10.1016/j.neucom.2021.07.098}
	353	+}
	354	+
	355	+@article{zhang2021flsurvey,
	356	+ author = {Zhang, Chen and Xie, Yu and Bai, Hang and Yu, Bin and Li, Weihong and Gao, Yuan},
	357	+ title = {{A survey on federated learning}},
	358	+ year = {2021},
	359	+ journal = {Knowledge-Based Systems},
	360	+ volume = {216},
	361	+ pages = {106775},
	362	+ doi = {10.1016/j.knosys.2021.106775}
	363	+}
	364	+
	365	+@article{xu2021flhealth,
	366	+ author = {Xu, Jie and Glicksberg, Benjamin S. and Su, Chang and Walker, Peter and Bian, Jiang and Wang, Fei},
	367	+ title = {{Federated Learning for Healthcare Informatics}},
	368	+ year = {2021},
	369	+ journal = {Journal of Healthcare Informatics Research},
	370	+ volume = {5},
	371	+ number = {1},
	372	+ pages = {1--19},
	373	+ doi = {10.1007/s41666-020-00082-4}
	374	+}
	375	+
	376	+@inproceedings{yang2024fedas,
	377	+ author = {Yang, Xiyuan and Huang, Wenke and Ye, Mang},
	378	+ title = {{FedAS: Bridging Inconsistency in Personalized Federated Learning}},
	379	+ year = {2024},
	380	+ booktitle = {2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
	381	+ pages = {11986--11995},
	382	+ publisher = {IEEE},
	383	+ doi = {10.1109/cvpr52733.2024.01139}
	384	+}
	385	+
	386	+@article{ali2026huntgpt,
	387	+ author = {Ali, Tarek and Kostakos, Panos and Sheikhi, Saeid},
	388	+ title = {{HuntGPT: Integrating Machine Learning-Based Anomaly Detection and Explainable AI with Large Language Models (LLMs)}},
	389	+ year = {2026},
	390	+ journal = {Telecom},
	391	+ volume = {7},
	392	+ number = {3},
	393	+ pages = {73},
	394	+ doi = {10.3390/telecom7030073}
	395	+}
	396	+
	397	+@inproceedings{bonawitz2017secagg,
	398	+ author = {Bonawitz, Keith and Ivanov, Vladimir and Kreuter, Ben and Marcedone, Antonio and McMahan, H. Brendan and Patel, Sarvar and Ramage, Daniel and Segal, Aaron and Seth, Karn},
	399	+ title = {{Practical Secure Aggregation for Privacy-Preserving Machine Learning}},
	400	+ year = {2017},
	401	+ booktitle = {Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security},
	402	+ pages = {1175--1191},
	403	+ publisher = {ACM},
	404	+ doi = {10.1145/3133956.3133982}
	405	+}
	406	+
	407	+@incollection{beullens2022rainbow,
	408	+ author = {Beullens, Ward},
	409	+ title = {{Breaking Rainbow Takes a Weekend on a Laptop}},
	410	+ year = {2022},
	411	+ booktitle = {Lecture Notes in Computer Science},
	412	+ pages = {464--479},
	413	+ publisher = {Springer Nature Switzerland},
	414	+ doi = {10.1007/978-3-031-15979-4_16}
	415	+}
	416	+
	417	+@article{uddin2026phishing,
	418	+ author = {Uddin, Mohammad Amaz and Mahiuddin, Md and Sarker, Iqbal H.},
	419	+ title = {{An explainable transformer-based model for phishing email detection: A large language model approach}},
	420	+ year = {2026},
	421	+ journal = {Computer Networks},
	422	+ volume = {277},
	423	+ pages = {112061},
	424	+ doi = {10.1016/j.comnet.2026.112061}
	425	+}
	426	+
	427	+@article{xu2025cbrfl,
	428	+ author = {Xu, Gang and Lei, Lele and Mao, Yanhui and Li, Zongpeng and Chen, Xiu-Bo and Zhang, Kejia},
	429	+ title = {{CBRFL: A framework for Committee-based Byzantine-Resilient Federated Learning}},
	430	+ year = {2025},
	431	+ journal = {Journal of Network and Computer Applications},
	432	+ volume = {238},
	433	+ pages = {104165},
	434	+ doi = {10.1016/j.jnca.2025.104165}
	435	+}
	436	+
	437	+@techreport{nist2024transition,
	438	+ author = {Regenscheid, Andrew},
	439	+ title = {{Transition to Post-Quantum Cryptography Standards}},
	440	+ year = {2024},
	441	+ institution = {National Institute of Standards and Technology},
	442	+ doi = {10.6028/NIST.IR.8547.ipd}
	443	+}
	444	+
	445	+@techreport{nist2025hqc,
	446	+ author = {Alagic, Gorjan and Bros, Maxime and Ciadoux, Pierre and Cooper, David and Dang, Quynh and Dang, Thinh and Kelsey, John and Lichtinger, Jacob and Liu, Yi-Kai and Miller, Carl and Moody, Dustin and Peralta, Rene and Perlner, Ray and Robinson, Angela and Silberg, Hamilton and Smith-Tone, Daniel and Waller, Noah},
	447	+ title = {{Status report on the fourth round of the NIST post-quantum cryptography standardization process}},
	448	+ year = {2025},
	449	+ institution = {National Institute of Standards and Technology (U.S.)},
	450	+ doi = {10.6028/NIST.IR.8545}
	451	+}
	452	+
	453	+@article{guo2022blockchainsecurity,
	454	+ author = {Guo, Huaqun and Yu, Xingjie},
	455	+ title = {{A survey on blockchain technology and its security}},
	456	+ year = {2022},
	457	+ journal = {Blockchain: Research and Applications},
	458	+ volume = {3},
	459	+ number = {2},
	460	+ pages = {100067},
	461	+ doi = {10.1016/j.bcra.2022.100067}
	462	+}
	463	+
	464	+@article{kim2018provenance,
	465	+ author = {Kim, Henry M. and Laskowski, Marek},
	466	+ title = {{Toward an ontology-driven blockchain design for supply-chain provenance}},
	467	+ year = {2018},
	468	+ journal = {Intelligent Systems in Accounting, Finance and Management},
	469	+ volume = {25},
	470	+ number = {1},
	471	+ pages = {18--27},
	472	+ doi = {10.1002/isaf.1424}
	473	+}
	474	+
	475	+@article{esposito2018healthcare,
	476	+ author = {Esposito, Christian and De Santis, Alfredo and Tortora, Genny and Chang, Henry and Choo, Kim-Kwang Raymond},
	477	+ title = {{Blockchain: A Panacea for Healthcare Cloud-Based Data Security and Privacy?}},
	478	+ year = {2018},
	479	+ journal = {IEEE Cloud Computing},
	480	+ volume = {5},
	481	+ number = {1},
	482	+ pages = {31--37},
	483	+ doi = {10.1109/mcc.2018.011791712}
	484	+}
	485	+
	486	+@misc{li2019qffl,
	487	+ author = {Tian Li and Maziar Sanjabi and Ahmad Beirami and Virginia Smith},
	488	+ title = {{Fair Resource Allocation in Federated Learning}},
	489	+ year = {2019},
	490	+ eprint = {1905.10497},
	491	+ archivePrefix = {arXiv},
	492	+ howpublished = {arXiv preprint arXiv:1905.10497}
	493	+}
	494	+
	495	+@misc{li2018fedprox,
	496	+ author = {Tian Li and Anit Kumar Sahu and Manzil Zaheer and Maziar Sanjabi and Ameet Talwalkar and Virginia Smith},
	497	+ title = {{Federated Optimization in Heterogeneous Networks}},
	498	+ year = {2018},
	499	+ eprint = {1812.06127},
	500	+ archivePrefix = {arXiv},
	501	+ howpublished = {arXiv preprint arXiv:1812.06127}
	502	+}
	503	+
	504	+@misc{wang2021nft,
	505	+ author = {Qin Wang and Rujia Li and Qi Wang and Shiping Chen},
	506	+ title = {{Non-Fungible Token (NFT): Overview, Evaluation, Opportunities and Challenges}},
	507	+ year = {2021},
	508	+ eprint = {2105.07447},
	509	+ archivePrefix = {arXiv},
	510	+ howpublished = {arXiv preprint arXiv:2105.07447}
	511	+}
	512	+
	513	+@misc{cho2020tracing,
	514	+ author = {Hyunghoon Cho and Daphne Ippolito and Yun William Yu},
	515	+ title = {{Contact Tracing Mobile Apps for COVID-19: Privacy Considerations and Related Trade-offs}},
	516	+ year = {2020},
	517	+ eprint = {2003.11511},
	518	+ archivePrefix = {arXiv},
	519	+ howpublished = {arXiv preprint arXiv:2003.11511}
	520	+}
	521	+
	522	+@misc{bagdasaryan2018backdoor,
	523	+ author = {Eugene Bagdasaryan and Andreas Veit and Yiqing Hua and Deborah Estrin and Vitaly Shmatikov},
	524	+ title = {{How To Backdoor Federated Learning}},
	525	+ year = {2018},
	526	+ eprint = {1807.00459},
	527	+ archivePrefix = {arXiv},
	528	+ howpublished = {arXiv preprint arXiv:1807.00459}
	529	+}
	530	+
	531	+@inproceedings{guesmi2026reentrancy,
	532	+ author = {Guesmi, Semia and Piazza, Carla and Gasparetto, Andrea and Rizzo, Matteo and Rossi, Sabina},
	533	+ title = {{Detecting Cross-Function Reentrancy from EVM Traces}},
	534	+ year = {2026},
	535	+ booktitle = {OASIcs, Volume 142, FMBC 2026 (6th International Workshop on Formal Methods for Blockchains)},
	536	+ pages = {8:1--8:15},
	537	+ publisher = {Schloss Dagstuhl -- Leibniz-Zentrum für Informatik},
	538	+ doi = {10.4230/oasics.fmbc.2026.8}
	539	+}

added manuscript/references_verification_part06.csv +63 −0

Large diff (63 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/references_verification_part07b.csv +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

modified manuscript/tables/TABLES.json +241 −0

Large diff (247 lines) is not loaded by default to keep the page fast.Load diff

modified manuscript/tables/numbers.tex +2302 −0

Large diff (2,305 lines) is not loaded by default to keep the page fast.Load diff

modified manuscript/tables/tab_cagr.tex +40 −40

Large diff (85 lines) is not loaded by default to keep the page fast.Load diff

modified manuscript/tables/tab_diversity.tex +8 −8

Large diff (22 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_ai_shares.tex +21 −0

Large diff (21 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_altdates.tex +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_area_flow_change.tex +19 −0

Large diff (19 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_area_shares.tex +22 −0

Large diff (22 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_category_periods.tex +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_classification.tex +20 −0

Large diff (20 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_cooccurrence.tex +26 −0

Large diff (26 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_cross_area.tex +22 −0

Large diff (22 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_emerging_cutoffs.tex +19 −0

Large diff (19 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_emerging_setup.tex +13 −0

Large diff (13 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_h3.tex +19 −0

Large diff (19 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_its_levels.tex +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_its_slopes.tex +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_its_total_events.tex +16 −0

Large diff (16 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_landmarks.tex +28 −0

Large diff (28 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_llm_partners.tex +26 −0

Large diff (26 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_modularity.tex +20 −0

Large diff (20 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_openness.tex +20 −0

Large diff (20 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_periods.tex +24 −0

Large diff (24 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_total_robustness.tex +23 −0

Large diff (23 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_p9_unknown_breaks.tex +54 −0

Large diff (54 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_security_counts.tex +23 −0

Large diff (23 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_security_landmarks.tex +16 −0

Large diff (16 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_security_llm_cooccurrence.tex +23 −0

Large diff (23 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_security_summary.tex +16 −0

Large diff (16 lines) is not loaded by default to keep the page fast.Load diff

added manuscript/tables/tab_systems_categories.tex +22 −0

Large diff (22 lines) is not loaded by default to keep the page fast.Load diff

modified results/bibliometrics/cagr_by_category.csv +40 −40

Large diff (81 lines) is not loaded by default to keep the page fast.Load diff

modified results/bibliometrics/category_share_ci_yearly.csv +320 −320

Large diff (641 lines) is not loaded by default to keep the page fast.Load diff

modified results/bibliometrics/diversity_yearly.csv +8 −8

Large diff (17 lines) is not loaded by default to keep the page fast.Load diff

modified tables/make_numbers.py +109 −0

Large diff (114 lines) is not loaded by default to keep the page fast.Load diff

modified tables/make_tables.py +287 −0

Large diff (299 lines) is not loaded by default to keep the page fast.Load diff