# REFERENCE_MAP — which verified references each manuscript part should cite (WP44, Agent B-lit, 2026-11-03)

Source of keys: `manuscript/references.bib` (every entry resolved against Crossref / OpenAlex / DataCite / arXiv or a
live URL; see `manuscript/references_verification.csv`). Keys are `surnameYYYYkeyword`. Entries marked **core** are the
minimum set the part needs; the others are supplementary and may be cited selectively. Group letters (A–J) refer to the
scope groups of `agents/briefs/AGENT_LIT_REFERENCES.md`. Nothing in this map asserts a finding of WP44; each line only
says what the reference *supports*. arXiv-only entries are preprints and must be described as such in the text.

Conventions: `key — why`. Parts follow CLAUDE.md §9 (I–XI + appendices).

## Part I — Introduction and Scientific Motivation

**core**
- openai2022chatgpt — the dated event (30 Nov 2022) that defines the research period; cite the archived announcement, not a secondary account.
- ouyang2022instructgpt — the instruction-tuned/RLHF recipe behind ChatGPT; establishes what was technically new.
- brown2020gpt3 — in-context learning at scale; the immediate pre-ChatGPT state of the art (baseline period).
- openai2023gpt4 — first major post-ChatGPT frontier release inside the window.
- bommasani2021foundation — "foundation model" vocabulary and the cross-field research agenda the manuscript tests empirically.
- fortunato2018scienceofscience ; wang2021scienceofscience — science-of-science framing: large-scale bibliographic data as the object of study.
- vannoorden2023survey ; messeri2024illusions ; birhane2023science — motivation: documented concern about AI changing how research is done.
- liang2024mapping ; kobak2025delving — evidence that LLM use in scientific writing rose sharply after 2022 (justifies studying the field's own output).
- krenn2023forecasting — AI literature growth modelled as an exponentially growing knowledge network; motivates forecasting/emergence questions.
- bornmann2021growth ; bornmann2015growth — long-run growth rates of science; needed to separate secular growth from a post-2022 break.
- park2023disruption ; chu2021slowed — the "science is slowing" debate the mapping must engage.
**supplementary**
- maslej2024aiindex ; maslej2025aiindex — descriptive statistics on AI publication/industry trends (secondary, cite as reports).
- ahmed2023industry — growing role of industry in AI research (motivates institution-type analyses).
- bianchini2022aiscience ; cockburn2018aiinnovation — AI as a general method of invention (economics framing).
- hicks2015leiden — responsible use of metrics; signals the manuscript's evaluative stance.
- dwivedi2023sowhat — multidisciplinary perspectives on ChatGPT for research and practice.

## Part II — The Evolution of Computer Science (conceptual and historical background)

**core (pre-2018 lineage of the deep-learning paradigm)**
- turing1950computing — origin of the machine-intelligence question.
- rumelhart1986backprop ; hochreiter1997lstm ; lecun2015deeplearning — backpropagation, recurrent memory, deep learning synthesis.
- krizhevsky2017alexnet ; he2016resnet — the 2012–2016 vision breakthroughs (CACM reprint of AlexNet is the verified DOI).
- goodfellow2014gan ; kingma2013vae — generative modelling before diffusion.
- mnih2015dqn ; silver2016alphago ; silver2017alphagozero ; silver2018alphazero ; schulman2017ppo ; christiano2017rlhf — deep RL lineage that leads to RLHF.
- sutskever2014seq2seq ; bahdanau2014attention ; vaswani2017attention — sequence modelling to the Transformer.
- devlin2019bert ; raffel2020t5 ; radford2019gpt2 ; brown2020gpt3 — pretrained language models 2018–2020 (baseline period).
- kaplan2020scaling ; hoffmann2022chinchilla ; sevilla2022compute ; thompson2020computationallimits — scaling laws and compute growth.
- shazeer2017moe ; fedus2022switch — sparse mixture-of-experts lineage.
- sutton2019bitterlesson — the "compute over priors" thesis often invoked to explain the paradigm shift (blog essay, cite as such).
**core (other CS areas' trajectories 2018–2022)**
- hennessy2019goldenage ; leiserson2020roomattop ; dennard1974scaling — end of Dennard scaling / domain-specific architecture narrative.
- jouppi2017tpu ; dally2021gpu ; sze2017efficientdnn — accelerator lineage.
- preskill2018nisq ; arute2019supremacy — quantum computing's NISQ era.
- shi2016edge ; satyanarayanan2017edge — edge computing.
- mckeown2008openflow ; bosshart2014p4 ; kreutz2015sdn — programmable networks.
- dean2008mapreduce ; zaharia2016spark ; corbett2013spanner ; verma2015borg ; burns2016kubernetes ; jonas2019serverless — cloud/distributed systems lineage.
- lamport1998paxos ; castro2002pbft ; gilbert2002cap — distributed consensus foundations; nakamoto2008bitcoin ; zheng2018blockchain — blockchain.
- gentry2009fhe ; dwork2006dp ; dwork2014dp ; shor1997factoring ; regev2009lwe ; bernstein2017pqc — cryptography/privacy lineage up to the PQC transition.
- leroy2009compcert ; klein2009sel4 ; jung2017rustbelt ; jung2021saferust ; matsakis2014rust — formal verification and memory-safe systems languages.
- cook1971complexity ; karp1972reducibility ; babai2016graphisomorphism ; raz2019bqpph ; ji2021mipre ; chen2022maxflow ; kelley2023threeap — landmark theory results framing "theory" as a continuing frontier.
- lattner2004llvm ; lattner2021mlir ; chen2018tvm ; paszke2019pytorch ; abadi2016tensorflow — compiler/framework infrastructure.
- mildenhall2020nerf ; karras2019stylegan ; ho2020ddpm ; rombach2022ldm ; radford2021clip ; dosovitskiy2021vit — vision/graphics turn to neural fields and diffusion.
- jumper2021alphafold ; baek2021rosettafold — AI-for-science landmark before the window.
- mcmahan2017fl ; kairouz2021fl — federated learning.
- amodei2016concrete ; hendrycks2021unsolved ; bender2021parrots — safety/ethics agenda before ChatGPT.
**supplementary**
- wu2015citeseerx ; giles1998citeseer ; ley2002dblp ; ginsparg2011arxiv — history of CS's own bibliographic infrastructure (arXiv, DBLP, CiteSeer).
- lariviere2014arxiv ; sutton2017arxivcs ; lin2020arxivcs — the rise of preprint culture in CS before 2022.
- frank2019citationgraphs ; klinger2020narrowing ; ahmed2020dedemocratization ; martinezplumed2021benchmarks — bibliometric portraits of AI research pre-2022.

## Part III — Data and Systematic Review Protocol

**core (sources — group A)**
- priem2022openalex ; openalex2024topics ; alperin2024openalex ; culbert2025openalexcoverage ; scheidsteger2023magopenalex — OpenAlex: description, topic classification, suitability and coverage evidence, MAG→OpenAlex continuity.
- ley2002dblp ; ley2009dblp — DBLP design, scope and curation lessons.
- ginsparg2011arxiv ; clement2019arxivdataset ; arxiv2026stats — arXiv as a source and its submission statistics.
- kinney2023s2ag ; lo2020s2orc ; ammar2018literaturegraph ; fricke2018semanticscholar — Semantic Scholar Academic Graph / S2ORC.
- hendricks2020crossref — Crossref metadata as the DOI backbone used for deduplication.
- radev2013aclanthology ; acl2024anthology — ACL Anthology.
- wang2020mag ; sinha2015mas — Microsoft Academic Graph (discontinued 2021; lineage of OpenAlex).
- giles1998citeseer ; wu2015citeseerx ; tang2008arnetminer ; herzog2020dimensions ; peroni2020opencitations — other sources considered.
- visser2021comparison ; martinmartin2021coverage — coverage comparisons between bibliographic databases (justify multi-source design).
**core (protocol — group B)**
- page2021prisma ; page2021prismaexplanation ; rethlefsen2021prismas ; haddaway2022prismaflow ; moher2009prisma — PRISMA 2020 reporting, search reporting (PRISMA-S), flow diagram.
- kitchenham2009slr ; kitchenham2013slrprocess ; kitchenham2015ebse — SE systematic-review guidelines (Kitchenham lineage; the 2007 EBSE technical report has no resolvable record, cite the 2015 book).
- petersen2008mapping ; petersen2015mapping — systematic mapping study method (the manuscript's design).
- tricco2018prismascr ; arksey2005scoping ; grant2009typology — scoping-review methodology and review typology.
- wohlin2014snowballing ; garousi2019multivocal — snowballing and grey-literature inclusion rules.
- elliott2014living ; elliott2017living — living-review design (daily increment pipeline).
- vandeschoot2021asreview ; omaraeide2015textmining ; marshall2019automation — screening automation (active learning).
- khraisha2024gpt4screening ; guo2024llmscreening ; syriani2023chatgptscreening ; dennstadt2024llmscreening ; alshami2023chatgptsr ; wang2024zeroshotscreening — evaluations of LLM-assisted screening 2023–2024 (justify and bound the use of LLM classifiers; report their error rates).
- zupic2015bibliometric ; donthu2021bibliometric — bibliometric-study design guidelines.
**core (integrity)**
- nosek2015top ; munafo2017manifesto ; wilkinson2016fair — open-science/TOP/FAIR standards the protocol follows.
- piwowar2018oa — OA prevalence/citation advantage (relevant to full-text access limits).

## Part IV — Computational Methodology

**core (indicators — group C)**
- waltman2016review — survey of citation-impact indicators (anchor for the indicator section).
- waltman2019fieldnormalization ; waltman2011crown ; purkayastha2019fwci — field normalisation (MNCS, FWCI).
- wang2013citationwindow — citation-window choice (crucial for 2023–2026 papers with short windows).
- hirsch2005hindex ; waltman2012hindex ; egghe2006gindex — h/g indices and their inconsistencies.
- nielsen2021inequality ; radicchi2008universality ; bornmann2008citations — citation inequality (Gini), universality, what citations measure.
- stirling2007diversity ; leinster2012diversity ; porter2009interdisciplinarity ; rafols2010diversity ; leydesdorff2019interdisciplinarity — Rao–Stirling diversity and interdisciplinarity measures.
- newman2001collaboration ; newman2004coauthorship ; katz1997collaboration ; wagner2017collaboration ; wagner2015cooperation ; lariviere2015teamsize ; wuchty2007teams — co-authorship network and team-size measures.
- price1965networks ; garfield1955citation — citation-network foundations.
- fire2019goodhart ; hicks2015leiden ; ioannidis2020citation ; sinatra2016impact ; fanelli2016publicationrate — caveats and author-level indicators.
**core (science mapping and embeddings — group D)**
- small1973cocitation ; kessler1963coupling ; boyack2010accuracy ; klavans2017taxonomy — co-citation, bibliographic coupling, which relation maps best.
- vaneck2010vosviewer ; vaneck2014visualizing ; waltman2010unified ; waltman2012classification ; subelj2016clustering ; sjogarde2018granularity ; chen2006citespace ; borner2003domains ; leydesdorff2009globalmap — mapping/clustering of publication networks and granularity choices.
- blondel2008louvain ; traag2019leiden ; newman2006modularity ; fortunato2010community — community detection (Leiden is the default).
- beltagy2019scibert ; cohan2020specter ; ostendorff2022scincl ; singh2023scirepeval ; reimers2019sbert ; gao2021simcse ; muennighoff2023mteb ; devlin2019bert — scientific-document embeddings and their benchmarks.
- mcinnes2018umap ; mcinnes2018umapjoss ; vandermaaten2008tsne — dimensionality reduction for semantic maps.
- campello2013hdbscan ; mcinnes2017hdbscan ; schubert2017dbscan — density clustering.
- blei2003lda ; blei2006dtm ; dieng2020etm ; grootendorst2022bertopic ; angelov2020top2vec ; greene2014stability ; roder2015coherence ; sievert2014ldavis — topic models, dynamic topics, stability and coherence diagnostics.
**core (structural change — group E)**
- bai1998breaks ; bai2003breaks ; andrews1993breaks ; andrews1994optimal ; chow1960test ; perron1989crash ; hansen2001structuralchange ; zeileis2002strucchange — multiple-break tests and dating (the 30 Nov 2022 break hypothesis).
- killick2012pelt ; fryzlewicz2014wbs ; frick2014smuce ; adams2007bocpd ; truong2020ruptures ; aminikhanghahi2017survey — change-point detection algorithms and reviews.
- newey1987hac — HAC standard errors for monthly series.
- hamilton1989regime — regime-switching alternative.
- bernal2017its ; bernal2018itscontrol ; kontopantelis2015its ; box1975intervention ; brodersen2015causalimpact — interrupted time series / intervention analysis with controls (other fields as controls).
- abadie2010synthetic ; abadie2021synthetic — synthetic control.
- sun2021eventstudy ; callaway2021did ; goodmanbacon2021did ; roth2022pretest ; roth2023trending — event-study/DiD pitfalls (staggered adoption, pre-trends).
- taylor2018prophet ; hyndman2018fpp — forecasting baselines and seasonality handling.
**core (emergence detection — group F)**
- rotolo2015emerging ; cozzens2010emerging ; small2014emergence ; wang2018emerging ; klavans2017prominence ; boyack2022forecasting ; guo2011mixedindicators ; shibata2008fronts ; porter2019emergence ; salatino2017topicsborn ; salatino2018augur ; small2006growth — definitions and operationalisations of emerging topics (the original-contribution section builds on these).
- kleinberg2003burst ; kleinberg2002burstkdd ; mane2004bursts — burst detection.
- ke2015sleepingbeauties — delayed recognition.
- funk2017disruption ; wu2019disruption ; park2023disruption ; bornmann2020disruptionvalidity ; leibel2024disruptionreview ; holst2024disruptionartefacts — disruption index, its validity and dataset-artefact critique.
- uzzi2013novelty ; wang2017novelty ; lee2015novelty ; foster2015tradition — novelty measures and bias against novelty.
- milojevic2015cognitiveextent ; bettencourt2009discovery ; chu2021slowed ; lin2023remote — cognitive extent, topological transitions, large-field slowdown.
**core (reproducibility of the pipeline — group I)**
- gundersen2018reproducibility ; pineau2021reproducibility ; kapoor2023leakage ; kapoor2024reforms ; semmelrock2025reproducibility ; gundersen2022sources ; dodge2019showyourwork ; gebru2021datasheets ; mitchell2019modelcards ; bender2018datastatements — standards applied to WP44's own classifiers and datasets.

## Part V — The Artificial Intelligence Transformation

**core (foundation models and training)**
- vaswani2017attention ; devlin2019bert ; brown2020gpt3 ; raffel2020t5 ; chowdhery2022palm ; openai2023gpt4 ; gemini2023 ; touvron2023llama ; touvron2023llama2 ; dubey2024llama3 ; deepseek2025r1 ; deepseek2025r1nature ; openai2024o1 — model lineage inside the window (open vs closed; reasoning models).
- kaplan2020scaling ; hoffmann2022chinchilla ; wei2022emergent ; schaeffer2023mirage — scaling and the emergent-abilities controversy (conflicting findings to be evaluated critically).
- ouyang2022instructgpt ; christiano2017rlhf ; bai2022hh ; bai2022constitutional ; rafailov2023dpo ; casper2023rlhfproblems — alignment training methods.
- fedus2022switch ; shazeer2017moe ; gu2023mamba ; dao2022flashattention ; hu2021lora ; dettmers2023qlora ; frantar2023gptq ; dettmers2022int8 ; leviathan2023speculative — efficiency: sparsity, state-space models, attention kernels, PEFT, quantisation, decoding.
- zhao2023llmsurvey ; naveed2023llmoverview ; minaee2024llmsurvey ; min2023plmsurvey ; wan2023efficientllm ; zhou2024inferencesurvey — LLM surveys (map the sub-literature).
**core (reasoning, retrieval, tools, agents)**
- wei2022cot ; kojima2022zeroshotcot ; yao2023tot ; shinn2023reflexion — reasoning prompting.
- lewis2020rag ; gao2023ragsurvey ; karpukhin2020dpr — retrieval augmentation.
- yao2023react ; schick2023toolformer ; qin2023toolllm ; mialon2023augmented — tool use.
- wang2024agentsurvey ; xi2023rise ; guo2024multiagent ; park2023generativeagents ; wu2023autogen ; jimenez2024swebench ; zhou2023webarena ; liu2023agentbench ; packer2023memgpt ; mei2024aios ; ge2023llmos — agents: surveys, systems, benchmarks, "LLM as OS".
**core (multimodal, vision, speech, generation)**
- radford2021clip ; dosovitskiy2021vit ; kirillov2023sam ; alayrac2022flamingo ; li2023blip2 ; liu2023llava ; yin2024mllmsurvey ; zhang2024mmllms ; baltrusaitis2019multimodal — multimodal models and surveys.
- ho2020ddpm ; rombach2022ldm ; saharia2022imagen ; ramesh2022dalle2 ; peebles2023dit ; esser2024sd3 — diffusion generation.
- radford2023whisper ; baevski2020wav2vec2 ; hsu2021hubert ; mehrish2023speechreview — speech.
**core (evaluation)**
- hendrycks2021mmlu ; wang2018glue ; srivastava2022bigbench ; liang2022helm ; chiang2024arena ; zheng2023judge ; chang2024evalsurvey — benchmarks and LLM-as-judge; sainz2023contamination ; ott2022benchmarks ; dehghani2021lottery ; raji2021everything ; kiela2021dynabench ; recht2019imagenet — contamination, saturation, benchmark lottery.
- ji2023hallucination ; huang2025hallucination — hallucination.
**core (safety and alignment)**
- hendrycks2021unsolved ; ji2023alignment ; amodei2016concrete ; weidinger2022risks ; bengio2024extremerisks ; bengio2025safetyreport ; anwar2024foundational ; hendrycks2023catastrophic ; shevlane2023extremerisks ; zou2023universal ; wei2023jailbroken ; bender2021parrots ; bubeck2023sparks — safety surveys, risk taxonomies, adversarial robustness; "Sparks" as a contested claim.
**core (code)**
- chen2021codex ; li2022alphacode ; nijkamp2023codegen ; austin2021mbpp ; roziere2023codellama ; li2023starcoder ; jiang2024codegensurvey ; fan2023llm4se ; hou2024llm4se ; xu2022codelms ; pearce2022asleep ; ziegler2022copilot ; peng2023copilot ; barke2023groundedcopilot ; liang2024usability ; vaithilingam2022expectation — code generation models, SE surveys, productivity and usability evidence.
**supplementary**
- maslej2024aiindex ; maslej2025aiindex — descriptive context.

## Part VI — Systems and Computing Infrastructure

**core**
- hennessy2019goldenage ; leiserson2020roomattop ; dennard1974scaling ; jouppi2017tpu ; jouppi2023tpuv4 ; reuther2022accelerators ; sze2017efficientdnn ; dally2021gpu ; gholami2024memorywall — architecture and accelerators.
- narayanan2021megatron ; shoeybi2019megatronlm ; rajbhandari2020zero ; zhao2023fsdp ; barham2022pathways ; abadi2016tensorflow ; paszke2019pytorch ; ansel2024pytorch2 ; kwon2023vllm ; pope2023inference ; sheng2023flexgen — training/inference systems for large models.
- lattner2021mlir ; lattner2004llvm ; chen2018tvm ; tillet2019triton ; cummins2023llmcompiler — compilers for ML and LLMs for compilers.
- patterson2021carbon ; patterson2022plateau ; strubell2019energy ; luccioni2023bloom ; luccioni2024powerhungry ; schwartz2020greenai ; henderson2020carbon ; masanet2020datacenters ; devries2023energy ; wu2022sustainableai — sustainable computing / energy and carbon.
- shi2016edge ; satyanarayanan2017edge ; zhou2019edgeintelligence ; deng2020edgeintelligence ; letaief2022edgeai6g ; dang2020sixg — edge computing and 6G.
- bosshart2014p4 ; mckeown2008openflow ; kreutz2015sdn ; hauser2023p4survey ; sapio2021switchml ; hoiland2018xdp ; vieira2020ebpf — programmable networks, in-network computing, XDP/eBPF.
- madhavapeddy2013unikernels ; kuenzer2021unikraft ; packer2023memgpt ; mei2024aios ; ge2023llmos — operating systems: unikernels and LLM-OS proposals.
- dean2008mapreduce ; zaharia2016spark ; corbett2013spanner ; verma2015borg ; burns2016kubernetes ; jonas2019serverless ; lamport1998paxos ; castro2002pbft ; gilbert2002cap — distributed systems canon.
- pan2024vectordb ; han2023vectordb ; taipalus2024vectordb ; johnson2021faiss ; malkov2020hnsw ; fernandez2023llmdata ; stonebraker2024goesaround ; li2021ai4db — databases: vector DBMS, ANN search, LLMs and data management.
- khattab2020colbert ; zhao2024denseretrieval ; robertson2009bm25 ; zhu2023llmir — information retrieval.
- mcmahan2017fl ; kairouz2021fl ; li2020flchallenges — federated learning as a systems topic.
- preskill2018nisq ; arute2019supremacy ; kim2023utility ; googlequantum2025belowthreshold ; cerezo2021vqa ; bharti2022nisq — quantum computing status 2018–2025.
- brohan2023rt2 ; openx2023embodiment ; firoozi2024fmrobotics ; hu2023generalrobots ; kim2024openvla ; black2024pi0 — robotics foundation models (cyber-physical systems).
- kerbl2023gaussian ; muller2022instantngp ; barron2021mipnerf ; mildenhall2020nerf ; poole2022dreamfusion ; chen2024gssurvey ; fei2024gssurvey ; tewari2022neuralrendering — graphics/neural rendering (placed here or in Part V as the taxonomy decides).

## Part VII — Theory, Algorithms, and Security

**core**
- cook1971complexity ; karp1972reducibility ; babai2016graphisomorphism ; raz2019bqpph ; ji2021mipre ; chen2022maxflow ; kelley2023threeap ; shor1997factoring ; regev2009lwe — theory landmarks and the quantum/lattice connection.
- demoura2021lean4 ; mathlib2020 ; yang2024formalmath ; first2023baldur ; leroy2009compcert ; klein2009sel4 ; jung2017rustbelt ; jung2021saferust ; matsakis2014rust — formal methods, Lean/mathlib, LLM-assisted proving, verified systems, Rust.
- nist2024fips203 ; nist2024fips204 ; nist2024fips205 ; bernstein2017pqc — post-quantum standards (2024) and rationale.
- gentry2009fhe ; brakerski2014bgv ; cheon2017ckks ; chillotti2020tfhe ; acar2018hesurvey ; marcolla2022fhesurvey ; groth2016snark ; thaler2022proofs ; evans2018mpc — FHE, ZK proofs, MPC.
- dwork2006dp ; dwork2014dp ; abadi2016dpsgd ; shokri2017membership ; carlini2021extracting ; carlini2024poisoning — privacy and ML privacy attacks.
- greshake2023injection ; perez2022ignore ; liu2023promptinjection ; shayegani2023adversarial ; wei2023jailbroken ; zou2023universal ; yao2024llmsecurity ; das2025llmsecurity ; fang2024oneday ; deng2023pentestgpt — LLM security: prompt injection, jailbreaks, surveys, offensive use.
- szegedy2014intriguing ; goodfellow2015fgsm ; kocher2019spectre ; lipp2018meltdown — adversarial examples and microarchitectural attacks (pre-window security landmarks).
- nakamoto2008bitcoin ; zheng2018blockchain ; monrat2019blockchain ; werner2022defi — blockchain/DeFi.

## Part VIII — Scientific Knowledge Networks

**core**
- newman2001collaboration ; newman2004coauthorship ; katz1997collaboration ; wuchty2007teams ; lariviere2015teamsize ; wagner2015cooperation ; wagner2017collaboration ; lin2023remote — co-authorship, teams, international collaboration.
- small1973cocitation ; kessler1963coupling ; boyack2010accuracy ; klavans2017taxonomy ; price1965networks ; garfield1955citation — citation-network constructs.
- blondel2008louvain ; traag2019leiden ; newman2006modularity ; fortunato2010community — community structure.
- stirling2007diversity ; leinster2012diversity ; porter2009interdisciplinarity ; rafols2010diversity ; leydesdorff2019interdisciplinarity — interdisciplinarity/diffusion measures.
- nielsen2021inequality ; radicchi2008universality ; sinatra2016impact ; ioannidis2020citation — inequality and author-level impact.
- ahmed2023industry ; ahmed2020dedemocratization ; klinger2020narrowing ; frank2019citationgraphs ; martinezplumed2021benchmarks — industry/academia, concentration and narrowing in AI.
- piwowar2018oa ; lin2020arxivcs ; lariviere2014arxiv ; fraser2020biorxiv ; fraser2021preprints ; sutton2017arxivcs — preprints and OA effects on diffusion.
- funk2017disruption ; wu2019disruption ; park2023disruption ; uzzi2013novelty ; ke2015sleepingbeauties — disruption/novelty/delayed recognition on the citation graph.
- osborne2024huggingface ; castano2023hfcarbon ; wolf2020transformers ; lhoest2021datasets ; longpre2024dataprovenance — software/dataset ecosystems as knowledge networks (Hugging Face).

## Part IX — Empirical Results

Cite methods references from Part IV when reporting each result; in addition:
- wang2013citationwindow ; waltman2019fieldnormalization — when presenting short-window and normalised citation results.
- bai2003breaks ; killick2012pelt ; truong2020ruptures ; bernal2017its ; abadie2021synthetic ; roth2023trending — when presenting break dates, ITS and control-field comparisons.
- liang2024mapping ; liang2024monitoring ; kobak2025delving ; geng2024writingstyle ; gray2024contamination ; latona2024reviewlottery — external estimates of LLM-modified text/reviews to compare against the corpus's own vocabulary shifts.
- ott2022benchmarks ; martinezplumed2021benchmarks — benchmark dynamics to compare with observed benchmark-paper growth.
- bornmann2021growth ; bornmann2015growth ; fanelli2016publicationrate — reference growth rates.
- maslej2024aiindex ; maslej2025aiindex ; arxiv2026stats — external descriptive series for triangulation (reports/statistics pages, not peer-reviewed).
- greene2014stability ; roder2015coherence ; subelj2016clustering — when reporting topic/cluster validation.

## Part X — The Next Research Frontiers

- wang2023aiscience ; abramson2024alphafold3 ; lin2023esm2 ; hayes2025esm3 ; watson2023rfdiffusion ; dauparas2022proteinmpnn ; nguyen2024evo ; cui2024scgpt ; kochkov2024neuralgcm ; bi2023pangu ; lam2023graphcast ; davies2021mathematics ; fawzi2022alphatensor ; degrave2022tokamak ; romeraparedes2024funsearch ; trinh2024alphageometry ; merchant2023gnome — AI for science frontier.
- lu2024aiscientist ; gottweis2025coscientist ; boiko2023coscientist ; si2024novelideas — autonomous research agents (hypotheses to test, not established findings).
- bengio2025safetyreport ; anwar2024foundational ; ji2023alignment — open safety problems.
- googlequantum2025belowthreshold ; kim2023utility ; nist2024fips203 — quantum/post-quantum frontier.
- yang2024formalmath ; first2023baldur — formal reasoning frontier.
- gu2023mamba ; deepseek2025r1 ; openai2024o1 — architectures and reasoning models.
- brohan2023rt2 ; black2024pi0 ; kim2024openvla — embodied AI.
- luccioni2024powerhungry ; devries2023energy — sustainability constraints.
- rotolo2015emerging ; boyack2022forecasting ; salatino2018augur ; krenn2023forecasting — how emergence/forecast claims should be stated.
- shneiderman2020hcai ; amershi2019guidelines ; weisz2024genaidesign ; zamfirescu2023johnny ; subramonyam2024gulf ; lee2022coauthor ; bansal2021wholeexceeds ; vaccaro2024humanai ; morris2023scientists — human–AI interaction frontier.

## Part XI — Discussion and Conclusion

- hicks2015leiden ; fire2019goodhart ; bornmann2008citations ; wang2017novelty — limits of citation-based evidence.
- messeri2024illusions ; birhane2023science ; dwivedi2023sowhat ; lund2023chatgpt ; stokelwalker2023author ; thorp2023notauthor ; nature2023groundrules ; else2023abstracts ; gao2023abstracts ; conroy2023paperfromscratch ; checco2021aipeerreview ; liang2024feedback — implications for scientific practice, authorship and peer review (liang2024feedback: large-scale evaluation of LLM-generated paper feedback).
- noy2023productivity ; brynjolfsson2023genaiwork ; dellacqua2023jagged ; eloundou2024gpts ; gao2024aibenefits ; hao2024aiexpands — productivity/impact evidence (external, mostly non-CS).
- park2023disruption ; holst2024disruptionartefacts ; leibel2024disruptionreview — why disruption-based conclusions must be hedged.
- visser2021comparison ; alperin2024openalex ; culbert2025openalexcoverage — source-coverage limitations.
- roth2022pretest ; sun2021eventstudy — why temporal association is not causation.

## Appendices

- Appendix (taxonomy): openalex2024topics ; waltman2012classification ; sjogarde2018granularity ; klavans2017taxonomy ; leydesdorff2009globalmap.
- Appendix (search strategies): rethlefsen2021prismas ; wohlin2014snowballing ; garousi2019multivocal ; haddaway2022prismaflow.
- Appendix (mathematical derivations): bai1998breaks ; andrews1993breaks ; newey1987hac ; killick2012pelt ; adams2007bocpd ; stirling2007diversity ; leinster2012diversity ; traag2019leiden ; mcinnes2018umap ; campello2013hdbscan ; funk2017disruption ; kleinberg2003burst.
- Appendix (robustness): wang2013citationwindow ; bornmann2020disruptionvalidity ; greene2014stability ; subelj2016clustering ; roth2023trending ; abadie2021synthetic.
- Appendix (data dictionaries): priem2022openalex ; kinney2023s2ag ; ley2009dblp ; hendricks2020crossref ; wilkinson2016fair ; gebru2021datasheets.
- Appendix (reproducibility): acm2020badging ; pineau2021reproducibility ; kapoor2024reforms ; nosek2015top ; munafo2017manifesto ; collberg2016repeatability ; stodden2018journalpolicy ; hutson2018crisis ; baker2016reproducibility ; henderson2018drlmatters ; lipton2019troubling ; sambasivan2021datawork ; paullada2021discontents ; koch2021datasetlife ; bommasani2023fmti ; magnusson2023olmo.

## Not available (verification failed — do not cite until a verifiable record is found)

- Kitchenham & Charters 2007 EBSE technical report (no DOI/stable URL) → use kitchenham2015ebse.
- McCarthy et al. 1955/2006 Dartmouth proposal (AI Magazine DOI not resolvable via Crossref/OpenAlex).
- Sculley et al. 2018 "Winner's Curse?" (OpenReview blocks automated access; HTTP 403).
- Papers with Code (site retired; redirects to Hugging Face).
- dblp statistics page (bot challenge); use ley2009dblp and the DBLP XML dump documentation instead.
- OpenAI "Sora" technical report page (403 / empty archived page).
- Haunschild 2024 OpenAlex affiliation-quality paper and a generic "LLMs for systematic review screening" item (titles could not be matched; placeholders dropped).