[
  {
    "id": "R01",
    "title": "Against All Likelihood",
    "authors": "David Lopez-Paz",
    "year": 2026,
    "url": "https://arxiv.org/abs/2611.09385v1",
    "review_depth": "Full uploaded paper.tex and aal.tex; online v1 HTML and selected PDF pages checked.",
    "contribution": "Connects truth-seeking to relevant invariance, discusses admissibility, and sketches multi-context LLM inference.",
    "limitation": "A position paper and algorithmic sketch, not a new evaluated CNS or LLM system. Published PDF date and arXiv submission date differ; see audit.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R02",
    "title": "Invariant Risk Minimization",
    "authors": "Martin Arjovsky; Leon Bottou; Ishaan Gulrajani; David Lopez-Paz",
    "year": 2019,
    "url": "https://arxiv.org/abs/1907.02893",
    "review_depth": "Primary abstract and formulation, also cross-checked against R01.",
    "contribution": "Precedent for seeking a representation with a common optimal predictor across environments.",
    "limitation": "Not a universal OOD guarantee; do not equate prompting over viewpoints with observed training environments.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R03",
    "title": "Causal inference using invariant prediction: identification and confidence intervals",
    "authors": "Jonas Peters; Peter Buhlmann; Nicolai Meinshausen",
    "year": 2016,
    "url": "https://arxiv.org/abs/1501.01332",
    "review_depth": "Primary text, invariant-prediction assumptions, identifiability discussion and proposition 1.",
    "contribution": "Precedent for testing mechanism stability under specified environmental changes.",
    "limitation": "Causal identification requires assumptions about interventions and the target mechanism; multiple compatible models can remain.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R04",
    "title": "The Risks of Invariant Risk Minimization",
    "authors": "Elan Rosenfeld; Pradeep Ravikumar; Andrej Risteski",
    "year": 2021,
    "url": "https://arxiv.org/abs/2010.05761",
    "review_depth": "Primary abstract and selected introduction/analysis.",
    "contribution": "Gives failure results for IRM in analyzed linear and nonlinear settings.",
    "limitation": "The results concern specified models, not a proof that all invariance-based methods fail.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R05",
    "title": "Does Invariant Risk Minimization Capture Invariance?",
    "authors": "Pritish Kamath; Akilesh Tangella; Danica Sutherland; Nathan Srebro",
    "year": 2021,
    "url": "https://proceedings.mlr.press/v130/kamath21a.html",
    "review_depth": "Publisher abstract and bibliographic record.",
    "contribution": "Shows gaps between practical linear IRM and intended invariance, including sampling sensitivity.",
    "limitation": "No transfer of these exact theorems to CNS is asserted; they motivate explicit negative controls.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R06",
    "title": "In Search of Lost Domain Generalization",
    "authors": "Ishaan Gulrajani; David Lopez-Paz",
    "year": 2021,
    "url": "https://arxiv.org/abs/2007.01434",
    "review_depth": "Primary abstract and benchmark framing.",
    "contribution": "Strong tuned ERM and controlled model selection are essential comparators.",
    "limitation": "Its benchmark is not narrative synthesis; exact comparative results are not generalized to CNS.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R07",
    "title": "Discovering Environments with XRM",
    "authors": "Mohammad Pezeshki; Diane Bouchacourt; Mark Ibrahim; Nicolas Ballas; Pascal Vincent; David Lopez-Paz",
    "year": 2024,
    "url": "https://proceedings.mlr.press/v235/pezeshki24a.html",
    "review_depth": "Publisher abstract and method summary.",
    "contribution": "Precedent for discovering environments rather than assuming annotated groups.",
    "limitation": "Not an off-the-shelf validator of admissible narrative transformations; proposed CNS use would require separate evaluation.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R08",
    "title": "Learning the Difference That Makes a Difference with Counterfactually-Augmented Data",
    "authors": "Divyansh Kaushik; Eduard Hovy; Zachary C. Lipton",
    "year": 2020,
    "url": "https://arxiv.org/abs/1909.12434",
    "review_depth": "Primary abstract and official ICLR/project record.",
    "contribution": "Precedent for contrastive examples whose minimal edits change labels.",
    "limitation": "Human counterfactual edits do not automatically correspond to causal interventions in the world.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R09",
    "title": "Beyond Accuracy: Behavioral Testing of NLP Models with CheckList",
    "authors": "Marco Tulio Ribeiro; Tongshuang Wu; Carlos Guestrin; Sameer Singh",
    "year": 2020,
    "url": "https://aclanthology.org/2020.acl-main.442/",
    "review_depth": "Official ACL abstract and bibliographic record.",
    "contribution": "Precedent for a capability-by-behavior test suite rather than aggregate accuracy alone.",
    "limitation": "The CNS transformation contracts and thresholds are proposed here; they are not results from this paper.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R10",
    "title": "Integrating Conflicting Data: The Role of Source Dependence",
    "authors": "Xin Luna Dong; Laure Berti-Equille; Divesh Srivastava",
    "year": 2009,
    "url": "https://www.vldb.org/pvldb/vol2/vldb09-pvldb47.pdf",
    "review_depth": "Primary abstract, introduction, assumptions and source-dependence sections; PDF pages 1 and 3 visually checked.",
    "contribution": "Direct precedent for accounting for copying when resolving conflicting records.",
    "limitation": "Its truth-discovery models have domain and dependence assumptions; shared correct facts alone do not prove copying.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R11",
    "title": "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks",
    "authors": "Patrick Lewis and colleagues",
    "year": 2020,
    "url": "https://arxiv.org/abs/2005.11401",
    "review_depth": "Primary abstract.",
    "contribution": "Necessary evidence-retrieval baseline and a route beyond parametric memory.",
    "limitation": "Retrieval can return wrong or dependent sources; successful retrieval is not verification.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R12",
    "title": "Self-Consistency Improves Chain of Thought Reasoning in Language Models",
    "authors": "Xuezhi Wang and colleagues",
    "year": 2023,
    "url": "https://arxiv.org/abs/2203.11171",
    "review_depth": "Primary abstract.",
    "contribution": "Necessary multiple-sample aggregation baseline.",
    "limitation": "Agreement across samples is not independent external evidence.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R13",
    "title": "Improving Factuality and Reasoning in Language Models through Multiagent Debate",
    "authors": "Yilun Du; Shuang Li; Antonio Torralba; Joshua B. Tenenbaum; Igor Mordatch",
    "year": 2024,
    "url": "https://proceedings.mlr.press/v235/du24e.html",
    "review_depth": "Publisher abstract and official project summary.",
    "contribution": "Positive precedent for dialectical multiagent procedures.",
    "limitation": "Needs matched-compute voting and single-agent controls before attributing gains to dialogue.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R14",
    "title": "Debate or Vote: Which Yields Better Decisions in Multi-Agent Large Language Models?",
    "authors": "Hyeong Kyu Choi; Xiaojin Zhu; Sharon Li",
    "year": 2025,
    "url": "https://arxiv.org/abs/2508.17536",
    "review_depth": "Primary abstract and selected argument/experimental framing.",
    "contribution": "Challenges the assumption that debate itself adds value beyond voting.",
    "limitation": "Its mathematical analysis has assumptions; it is not a universal impossibility result for evidence-acquiring agents.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R15",
    "title": "Multi-Agent Debate for LLM Judges with Adaptive Stability Detection",
    "authors": "Tianyu Hu; Zhen Tan; Song Wang; Huaizhi Qu; Tianlong Chen",
    "year": 2025,
    "url": "https://proceedings.neurips.cc/paper_files/paper/2025/hash/42475c537936b2394b5015e871765056-Abstract-Conference.html",
    "review_depth": "Official NeurIPS abstract.",
    "contribution": "Counterpoint reporting benefits from adaptive debate in judgment tasks.",
    "limitation": "Judgment accuracy and modeled consensus dynamics do not establish truth of newly synthesized scientific accounts.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R16",
    "title": "Chain-of-Verification Reduces Hallucination in Large Language Models",
    "authors": "Shehzaad Dhuliawala and colleagues",
    "year": 2023,
    "url": "https://arxiv.org/abs/2309.11495",
    "review_depth": "Primary abstract.",
    "contribution": "Precedent for separately generated verification questions and answer checking.",
    "limitation": "Verification without independent evidence can retain common-mode errors; include a strong evidence-enabled variant.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R17",
    "title": "Large Language Models Cannot Self-Correct Reasoning Yet",
    "authors": "Jie Huang and colleagues",
    "year": 2024,
    "url": "https://arxiv.org/abs/2310.01798",
    "review_depth": "Primary abstract and official ICLR/DeepMind record.",
    "contribution": "Negative evidence against treating unaided self-critique as reliable correction.",
    "limitation": "A statement about studied models and protocols, not a claim about every future model or externally grounded correction.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R18",
    "title": "Language Models Do Not Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting",
    "authors": "Miles Turpin; Julian Michael; Ethan Perez; Samuel R. Bowman",
    "year": 2023,
    "url": "https://arxiv.org/abs/2305.04388",
    "review_depth": "Primary abstract.",
    "contribution": "Motivates observable evidence traces and executable predictions instead of trusting narrated rationales.",
    "limitation": "Does not imply every explanation is false; it prevents treating an explanation as verified internal causation.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R19",
    "title": "SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models",
    "authors": "Potsawee Manakul; Adian Liusie; Mark J. F. Gales",
    "year": 2023,
    "url": "https://arxiv.org/abs/2303.08896",
    "review_depth": "Primary abstract.",
    "contribution": "Useful consistency-based hallucination-detection comparator.",
    "limitation": "A consistency signal can miss stable common-mode falsehoods.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R20",
    "title": "Modern Bayesian Experimental Design",
    "authors": "Tom Rainforth; Adam Foster; Desi R. Ivanova; Freddie Bickford Smith",
    "year": 2024,
    "url": "https://arxiv.org/abs/2302.14545",
    "review_depth": "Primary overview and information-theoretic/adaptive design sections.",
    "contribution": "Established basis for choosing experiments by expected information or utility.",
    "limitation": "A predictive model or simulator is needed; a heuristic disagreement score must not be mislabeled exact information gain.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R21",
    "title": "Inductive Logic Programming at 30: A New Introduction",
    "authors": "Andrew Cropper; Sebastijan Dumancic",
    "year": 2022,
    "url": "https://arxiv.org/abs/2008.07912",
    "review_depth": "Primary overview and selected discussion of learned logic programs.",
    "contribution": "Precedent for bounded hypothesis languages, explicit rules and predicate invention.",
    "limitation": "CNS extraction and noisy evidence do not inherit soundness or completeness merely by using logic syntax.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R22",
    "title": "Strictly Proper Scoring Rules, Prediction, and Estimation",
    "authors": "Tilmann Gneiting; Adrian E. Raftery",
    "year": 2007,
    "url": "https://sites.stat.washington.edu/raftery/Research/PDF/Gneiting2007jasa.pdf",
    "review_depth": "Primary abstract, opening definitions and PDF first-page check.",
    "contribution": "Established foundation for honest probabilistic forecasting and proper scores.",
    "limitation": "Proper scoring incentivizes a truthful stated distribution in expectation; it does not make a model omniscient or guarantee calibration under arbitrary shift.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R23",
    "title": "Selective Classification for Deep Neural Networks",
    "authors": "Yonatan Geifman; Ran El-Yaniv",
    "year": 2017,
    "url": "https://arxiv.org/abs/1705.08500",
    "review_depth": "Primary abstract.",
    "contribution": "Precedent for evaluating the tradeoff between error and abstention.",
    "limitation": "CNS must assess coverage and error jointly; its coverage policy requires its own calibration.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R24",
    "title": "Conformal Risk Control",
    "authors": "Anastasios N. Angelopoulos; Stephen Bates; Adam Fisch; Lihua Lei; Tal Schuster",
    "year": 2024,
    "url": "https://arxiv.org/abs/2208.02814",
    "review_depth": "Primary overview and assumptions framing.",
    "contribution": "Optional precedent for calibration of bounded monotone losses with finite-sample control.",
    "limitation": "No unconditional guarantee under arbitrary distribution shift is imported; optional component, not required for TRACE.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R25",
    "title": "DiscoveryBench: Towards Data-Driven Discovery with Large Language Models",
    "authors": "Bodhisattwa Prasad Majumder and colleagues",
    "year": 2025,
    "url": "https://arxiv.org/abs/2407.01725",
    "review_depth": "Primary text overview and benchmark/task construction.",
    "contribution": "External discovery-oriented evaluation, complementing claim verification.",
    "limitation": "Existing benchmark tasks may be contaminated; their findings are not evidence that TRACE works.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R26",
    "title": "FEVER: a large-scale dataset for Fact Extraction and VERification",
    "authors": "James Thorne; Andreas Vlachos; Christos Christodoulopoulos; Arpit Mittal",
    "year": 2018,
    "url": "https://arxiv.org/abs/1803.05355",
    "review_depth": "Primary abstract and metadata.",
    "contribution": "Large evidence-grounded supported/refuted/not-enough-information benchmark.",
    "limitation": "Wikipedia-derived textual verification is not causal discovery or a direct world-truth oracle.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R27",
    "title": "Fact or Fiction: Verifying Scientific Claims",
    "authors": "David Wadden and colleagues",
    "year": 2020,
    "url": "https://aclanthology.org/2020.emnlp-main.609/",
    "review_depth": "Official ACL abstract and task description.",
    "contribution": "SciFact supplies scientific claims linked to annotated abstract evidence.",
    "limitation": "An abstract-entailment label does not independently replicate a scientific finding.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R28",
    "title": "AVeriTeC: A Dataset for Real-world Claim Verification with Evidence from the Web",
    "authors": "Michael Sejr Schlichtkrull; Zhijiang Guo; Andreas Vlachos",
    "year": 2023,
    "url": "https://fever.ai/dataset/averitec.html",
    "review_depth": "Primary paper abstract and current official dataset documentation.",
    "contribution": "Real-world claims with question-answer evidence and an explicit conflicting-evidence label.",
    "limitation": "Pin the dataset release, evidence snapshot and label date; do not silently mix later shared-task versions.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R29",
    "title": "Tensor Logic: The Language of AI",
    "authors": "Pedro Domingos",
    "year": 2025,
    "url": "https://arxiv.org/abs/2510.12269",
    "review_depth": "Primary abstract, introduction and Datalog background.",
    "contribution": "Actual precedent for representing symbolic computation in tensor form.",
    "limitation": "A language proposal does not establish that tensor geometry improves CNS truth-finding; benchmarking is still required.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R30",
    "title": "In-Context Symmetries: Self-Supervised Learning through Contextual World Models",
    "authors": "Sharut Gupta; Chenyu Wang; Yifei Wang; Tommi Jaakkola; Stefanie Jegelka",
    "year": 2024,
    "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/bcad07d4bfab51243efaa08b8ed475b3-Abstract-Conference.html",
    "review_depth": "Official primary abstract.",
    "contribution": "Additional precedent for context-dependent invariance and equivariance.",
    "limitation": "Transfer from its representation-learning setting to narrative truth assessment is an analogy to test, not established evidence.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G01",
    "title": "CNS 8.0 / Grounded Dialectical Orthesis",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/",
    "review_depth": "Relevant page text reviewed; Hub and current framing.",
    "contribution": "Hub and current framing.",
    "limitation": "Research/implementation/test proposal, not proof of a working production system.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G02",
    "title": "CNS 8.0 Theory",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/theory/",
    "review_depth": "Relevant page text reviewed; SNOs, chirality, entanglement, synthesis and orthesis definitions.",
    "contribution": "SNOs, chirality, entanglement, synthesis and orthesis definitions.",
    "limitation": "Definitions and research framing are not measured validation.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G03",
    "title": "CNS 8.0 Theory Claims and Assumptions",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/theory-claims-assumptions/",
    "review_depth": "Relevant page text reviewed; Assumptions and testable claims, including limits of topological diagnostics.",
    "contribution": "Assumptions and testable claims, including limits of topological diagnostics.",
    "limitation": "Existing caveats matter; do not portray CNS 8.0 as asserting metaphysical certainty.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G04",
    "title": "CNS 8.0 Mathematical Specification",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/mathematical-specification/",
    "review_depth": "Relevant page text reviewed; Closure, evidence gates, bundle/holonomy and acceptance formulations.",
    "contribution": "Closure, evidence gates, bundle/holonomy and acceptance formulations.",
    "limitation": "Derivability from grounded premises is different from truth of those premises.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G05",
    "title": "CNS Source Lineage Matrix",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/source-lineage-matrix/",
    "review_depth": "Relevant page text reviewed; Tracks ideas across CNS 2 through 8 and GCTS.",
    "contribution": "Tracks ideas across CNS 2 through 8 and GCTS.",
    "limitation": "An index of lineage, not an independent evaluation of each historical implementation.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G06",
    "title": "CNS 8.0 Experiments",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/experiments/",
    "review_depth": "Relevant page text reviewed; Existing proposed experiments and ablation directions.",
    "contribution": "Existing proposed experiments and ablation directions.",
    "limitation": "Plans and targets are not experimental outcomes.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G07",
    "title": "CNS 8.0 Metrics and Acceptance Criteria",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/metrics-acceptance-criteria/",
    "review_depth": "Relevant page text reviewed; Formal support, citation and synthesis metrics.",
    "contribution": "Formal support, citation and synthesis metrics.",
    "limitation": "Zero proof-trace failure does not imply zero false real-world claims.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G08",
    "title": "CNS 8.0 Oracle Boundary",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/oracle-boundary/",
    "review_depth": "Relevant page text reviewed; Separates runtime from held-out labels and evaluator information.",
    "contribution": "Separates runtime from held-out labels and evaluator information.",
    "limitation": "A sound boundary still needs enforced process and data separation.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G09",
    "title": "Worked Example - CNS 8.0 Resolves a Conditional Contradiction",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/worked-example/",
    "review_depth": "Relevant page text reviewed; Illustrates dose-context synthesis and residual uncertainty.",
    "contribution": "Illustrates dose-context synthesis and residual uncertainty.",
    "limitation": "Illustrative example, not a clinical finding or empirical result; same endpoint and same evidence record must be separated.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G10",
    "title": "CNS Tensor Logic and Predicate Invention",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns/tensor-logic-predicate-invention/",
    "review_depth": "Relevant page text reviewed; Links strict proof and exploratory latent-context proposals.",
    "contribution": "Links strict proof and exploratory latent-context proposals.",
    "limitation": "Proposed predicates require independent measurable validation, not merely lower residual contradiction.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G11",
    "title": "Grounded Chiral Tensor Synthesis Theory",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns-gcts/theory/",
    "review_depth": "Relevant page text reviewed; Grounded facts, candidate worlds, uncertainty and strict/probabilistic separation.",
    "contribution": "Grounded facts, candidate worlds, uncertainty and strict/probabilistic separation.",
    "limitation": "Likely-world ranking remains conditional on evidence and access assumptions.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G12",
    "title": "GCTS Record-Access Ontology",
    "authors": "GTCode Editorial / GTCode.com",
    "year": 2026,
    "url": "https://gtcode.com/guides/cns-gcts/record-access-ontology/",
    "review_depth": "Relevant page text reviewed; Explicitly models missing, available and restricted records.",
    "contribution": "Explicitly models missing, available and restricted records.",
    "limitation": "Missingness is not automatic evidence of wrongdoing, deception or a chosen hypothesis.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G13",
    "title": "CNS 2.0 Research Roadmap: Blueprint",
    "authors": "Ekewaka Lono",
    "year": 2025,
    "url": "https://gtcode.com/guides/cns-2.0-research-roadmap/blueprint/",
    "review_depth": "Relevant page text reviewed; Historical SNO/critics/entanglement/synthesis blueprint.",
    "contribution": "Historical SNO/critics/entanglement/synthesis blueprint.",
    "limitation": "An older current-version banner can lag the CNS 8.0 hub.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G14",
    "title": "Geometric Semantic Communication for Emergent Multi-Agent Intelligence, Appendix C",
    "authors": "Ekewaka Lono",
    "year": 2025,
    "url": "https://gtcode.com/papers/geometric-semantic-communication/",
    "review_depth": "Relevant page text reviewed; Appendix C perspective manifold and chiral synthesis formalism; surrounding tensor appendix inspected.",
    "contribution": "Appendix C perspective manifold and chiral synthesis formalism; surrounding tensor appendix inspected.",
    "limitation": "Inspection focused on CNS-relevant sections, not every section of the larger paper.",
    "accessed": "2026-11-15"
  },
  {
    "id": "G15",
    "title": "CNS 2.0: A Practical Blueprint for Chiral Narrative Synthesis",
    "authors": "Paul Lowndes",
    "year": 2025,
    "url": "https://gtcode.com/papers/ResearchProposal-ChiralNarrativeSynthesis_20250617_3.pdf",
    "review_depth": "Relevant page text reviewed; Eight-page original proposal read in parsed form; PDF pages 1 and 4 visually checked.",
    "contribution": "Eight-page original proposal read in parsed form; PDF pages 1 and 4 visually checked.",
    "limitation": "Title page credits Paul Lowndes; do not silently substitute the website author identity. Proposal rather than reported validation.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R31",
    "title": "Active Learning for Structure in Bayesian Networks",
    "authors": "Simon Tong; Daphne Koller",
    "year": 2001,
    "url": "https://ijcai.org/proceedings/2001/118",
    "review_depth": "Primary IJCAI proceedings page, abstract, metadata and BibTeX inspected; not full text.",
    "contribution": "Precedent for choosing interventions that are informative for causal structure learning.",
    "limitation": "Assumes a formal Bayesian-network learning setting; does not validate language-mediated CNS discovery.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R32",
    "title": "On the acceptability of arguments and its fundamental role in nonmonotonic reasoning, logic programming and n-person games",
    "authors": "Phan Minh Dung",
    "year": 1995,
    "url": "https://doi.org/10.1016/0004-3702(94)00041-X",
    "review_depth": "Publisher article abstract and metadata inspected; not full article.",
    "contribution": "Foundational abstract argumentation framework for conflicts between arguments.",
    "limitation": "Formal acceptability is not direct external correspondence to the world.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R33",
    "title": "ILP-based Inference for Cost-based Abduction on First-order Predicate Logic",
    "authors": "Naoya Inoue; Kentaro Inui",
    "year": 2013,
    "url": "https://doi.org/10.5715/jnlp.20.629",
    "review_depth": "Journal abstract, authors, DOI and issue metadata inspected; not full text.",
    "contribution": "Established computational precedent for cost-based abductive explanation search.",
    "limitation": "Tractable logic abduction does not by itself guarantee an explanation is empirically correct.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R34",
    "title": "The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery",
    "authors": "Chris Lu; Cong Lu; Robert Tjarko Lange; Jakob Foerster; Jeff Clune; David Ha",
    "year": 2024,
    "url": "https://arxiv.org/abs/2408.06292",
    "review_depth": "Primary arXiv abstract and metadata reviewed; not full manuscript.",
    "contribution": "Illustrates existing agentic end-to-end scientific idea, coding, experimentation and writing workflows.",
    "limitation": "Its task, benchmark and evaluation methodology differ from CNS proposed contrastive synthesis.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R35",
    "title": "Accelerating scientific discovery with Co-Scientist",
    "authors": "Juraj Gottweis; Wei-Hung Weng; Alexander Daryin; Tao Tu; et al.",
    "year": 2026,
    "url": "https://doi.org/10.1038/s41586-026-10644-y",
    "review_depth": "Primary Nature article landing page, abstract, bibliography information and available explanatory passages inspected; not full article.",
    "contribution": "Recent scientific hypothesis generation and agentic critique/revision system; strong related-work boundary.",
    "limitation": "Does not establish CNS-specific predictive-synthesis advantage or test its transformation contracts.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R36",
    "title": "A multi-agent system for automating scientific discovery",
    "authors": "Ali E. Ghareeb; Benjamin Chang; Ludovico Mitchener; et al.",
    "year": 2026,
    "url": "https://doi.org/10.1038/s41586-026-10652-y",
    "review_depth": "Primary Nature article landing page, abstract, method overview and bibliography information inspected; not full article.",
    "contribution": "Robin connects literature search, hypothesis generation and data analysis with experimental biology.",
    "limitation": "Its biological task setting and evidence differ from the proposed CNS benchmark and claims.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R37",
    "title": "Hypothesis Generation with Large Language Models",
    "authors": "Yangqiaoyu Zhou; Haokun Liu; Tejes Srivastava; Hongyuan Mei; Chenhao Tan",
    "year": 2024,
    "url": "https://arxiv.org/abs/2404.04326v3",
    "review_depth": "Primary v3 HTML introduction and method sections inspected; no code reproduction or full experimental audit.",
    "contribution": "Iterative hypothesis-bank updates from labeled prediction errors; a close H2 comparator precedent.",
    "limitation": "Classification-based hypothesis utility is not independent causal discovery or evaluation of CNS.",
    "accessed": "2026-11-15",
    "venue": "Proceedings of the 1st Workshop on NLP for Science (NLP4Science), 2024, pp. 117–139; DOI 10.18653/v1/2024.nlp4science-1.10"
  },
  {
    "id": "R38",
    "title": "Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement",
    "authors": "Linlu Qiu; Liwei Jiang; Ximing Lu; Melanie Sclar; Valentina Pyatkin; Chandra Bhagavatula; Bailin Wang; Yoon Kim; Yejin Choi; Nouha Dziri; Xiang Ren",
    "year": 2024,
    "url": "https://arxiv.org/abs/2310.08559v4",
    "review_depth": "Primary v4 HTML abstract and experimental-method framing inspected; not all appendices or code.",
    "contribution": "Proposing, selecting and refining rules with symbolic interpretation; distinguishes induction from application.",
    "limitation": "Task-specific rule interpreters and benchmarks do not directly implement provenance-aware CNS.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R39",
    "title": "Causal Modelling Agents: Causal Graph Discovery Through Synergising Metadata- and Data-Driven Reasoning",
    "authors": "Ahmed Abdulaal; Adamos Hadjivasiliou; Nina Montana-Brown; Tiantian He; Ayodeji Ijishakin; Ivana Drobnjak; Daniel C. Castro; Daniel C. Alexander",
    "year": 2024,
    "url": "https://discovery.ucl.ac.uk/id/eprint/10195820/1/8513_Causal_Modelling_Agents_C.pdf",
    "review_depth": "Author institutional ICLR 2024 PDF identity and opening methodological description inspected; no complete method/code replication.",
    "contribution": "Hybrid LLM metadata reasoning and data-driven structural causal modeling.",
    "limitation": "Does not establish the CNS composition or a general active-intervention capability; the unrelated arXiv:2402.09626 is not this work.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R40",
    "title": "Doing Experiments and Revising Rules with Natural Language and Probabilistic Reasoning",
    "authors": "Wasu Top Piriyakulkij; Cassidy Langenfeld; Tuan Anh Le; Kevin Ellis",
    "year": 2024,
    "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5f1b350fc0c2affd56f465faa36be343-Abstract-Conference.html",
    "review_depth": "Official NeurIPS 2024 abstract and opening paper description reviewed; no implementation reproduction.",
    "contribution": "Combines LLM natural-language rule proposals with probabilistic online updates and experiment design; close precedent for the investigation loop.",
    "limitation": "Studied rule-inference tasks do not establish the value of TRACE conflict alignment or source-lineage accounting.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R41",
    "title": "Contrastive Discovery: Open-Ended Scientific Discovery over Competing Explanations",
    "authors": "Ziang Liu; James J. Kim; Yijia Dai; Jennifer J. Sun",
    "year": 2026,
    "url": "https://arxiv.org/abs/2611.14514v1",
    "review_depth": "Primary HTML introduction, contrast-set/belief/search methods, evaluation framing, dataset and mechanical-audit appendices; author project page. Code not reproduced.",
    "contribution": "Explicit contrast sets and evidence-guided discovery; direct comparative precedent.",
    "limitation": "Belief concentration and expert agreement do not certify new externally true conditions. Project page lists code as soon; adaptation fidelity must be documented.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R42",
    "title": "Automated Hypothesis Validation with Agentic Sequential Falsifications",
    "authors": "Kexin Huang; Ying Jin; Ryan Li; Michael Y. Li; Emmanuel Candes; Jure Leskovec",
    "year": 2025,
    "url": "https://arxiv.org/abs/2502.09858v1",
    "review_depth": "Primary method, assumptions and sequential-validity argument; official repository entry point and sequential-testing paths at 59611864e09dd15d7528e240ad87c7ddbb4a701c; not executed.",
    "contribution": "POPPER: adaptive falsification with statistical error control under stated assumptions.",
    "limitation": "Guarantees depend on implication and conditional validity; arbitrary reused data or LLM scores do not meet them automatically.",
    "accessed": "2026-11-15",
    "implementation_url": "https://github.com/snap-stanford/POPPER/tree/59611864e09dd15d7528e240ad87c7ddbb4a701c"
  },
  {
    "id": "R43",
    "title": "Same Claim, Different Interpretations: How LLM Agents Resolve Ambiguity in Scientific Claims",
    "authors": "Delip Rao; Ahmed Muharram; Chris Callison-Burch",
    "year": 2026,
    "url": "https://arxiv.org/abs/2610.20457v1",
    "review_depth": "Primary HTML introduction, card mechanism, experimental framing and selected consistency appendix.",
    "contribution": "Claim Cards makes interpretations explicit before evidence collection.",
    "limitation": "Interpretation logging does not by itself improve agreement or demonstrate predictive discovery.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R44",
    "title": "Benchmarking LLM’s Capability in Reasoning over Conflicting Web References",
    "authors": "Yizhen Yuan; Rui Kong; Dongze Li; Yuanchun Li; Yunxin Liu",
    "year": 2026,
    "url": "https://aclanthology.org/2026.acl-long.11/",
    "review_depth": "Official ACL metadata; downloaded PDF introduction, benchmark/task construction and selected evaluation/limitations.",
    "contribution": "ConfRAG evaluates organization and explanation of conflicting web evidence.",
    "limitation": "Conflict-reasoning labels are not prospective prediction outcomes or a source-independence certificate.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R45",
    "title": "Failing to Falsify: Evaluating and Mitigating Confirmation Bias in Language Models",
    "authors": "Ayush Rajesh Jhaveri; Anthony GX-Chen; Ilia Sucholutsky; Eunsol Choi",
    "year": 2026,
    "url": "https://arxiv.org/abs/2604.02485v2",
    "review_depth": "Primary v2 abstract, selected Wason/Blicket results and evaluation framing, version metadata and method/appendix navigation; detailed statistical implementation not audited.",
    "contribution": "Simple counterexample-oriented prompting is a required low-complexity control.",
    "limitation": "Hidden-rule tasks do not establish source-conflict discovery or general scientific validity.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R46",
    "title": "Perceive-Then-Adjudicate: A Multi-Agent Framework for Resolving Knowledge-Evidence Conflicts in LLM Fact-Checking",
    "authors": "Yongcheng Zhang; Xin Zhang; Fei Li; Chong Teng; Jiang-Ming Yang; Donghong Ji; Zhuang Li",
    "year": 2026,
    "url": "https://aclanthology.org/2026.findings-emnlp.1445/",
    "review_depth": "Official ACL record; downloaded PDF opening, architecture and selected experimental framing.",
    "contribution": "Knowledge/evidence arbitration and rejection of erroneous consensus.",
    "limitation": "Fact-checking performance is not held-out explanatory-condition discovery; code link announced, not executed.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R47",
    "title": "Perspective—Discovery Within Validation Logic: Deliberately Surfacing, Complementing, and Substituting Abductive Reasoning in Hypothetico-Deductive Inquiry",
    "authors": "Kristin Behfar; Gerardo A. Okhuysen",
    "year": 2018,
    "url": "https://doi.org/10.1287/orsc.2017.1193",
    "review_depth": "Publisher full HTML selected sections on abductive anomaly/discordance and theory development; metadata.",
    "contribution": "Contradictory findings as an impetus for abductive explanation predates TRACE.",
    "limitation": "Methodological perspective, not a computational benchmark result.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R48",
    "title": "PARADOX: pharmacological analysis of relation anomalies and drug contradictions extraction",
    "authors": "Ali Fata; Hossein Rahmani",
    "year": 2026,
    "url": "https://doi.org/10.1007/s10822-026-00965-5",
    "review_depth": "Publisher abstract, metadata and data availability; selected cells in all three official notebooks and supplied CSV headers/counts at 10d7a04d50e72dc375cff4146e19a42c5b022dec. Subscription full article unavailable; code not executed.",
    "contribution": "Contradiction analysis includes condition differences and missing context.",
    "limitation": "Inspected evaluation uses GPT reference labels. No committed held-out condition-prediction workflow identified in inspected notebooks; full article not reviewed, so operational differentiation remains unresolved.",
    "accessed": "2026-11-15",
    "implementation_url": "https://github.com/ali-ft/PARADOX/tree/10d7a04d50e72dc375cff4146e19a42c5b022dec"
  },
  {
    "id": "R49",
    "title": "Utilizing ChatGPT as a scientific reasoning engine to differentiate conflicting evidence and summarize challenges in controversial clinical questions",
    "authors": "Shiyao Xie and colleagues",
    "year": 2024,
    "url": "https://doi.org/10.1093/jamia/ocae100",
    "review_depth": "Primary PMC full article: abstract, evidence/assertion workflow and selected error discussion.",
    "contribution": "Earlier LLM-assisted differentiation of conflicting scientific evidence.",
    "limitation": "Evidence synthesis and expert assessment do not establish the proposed sealed predictive endpoint.",
    "accessed": "2026-11-15"
  },
  {
    "id": "R50",
    "title": "AutoDiscovery: Open-ended Scientific Discovery via Bayesian Surprise",
    "authors": "Dhruv Agarwal; Bodhisattwa Prasad Majumder; Reece Adamson; Megha Chakravorty; Satvika Reddy Gavireddy; Aditya Parashar; Harshit Surana; Bhavana Dalvi Mishra; Andrew McCallum; Ashish Sabharwal; Peter Clark",
    "year": 2025,
    "url": "https://proceedings.neurips.cc/paper_files/paper/2025/hash/23b127521af7ca7a42f5cdb7507be4f2-Abstract-Conference.html",
    "review_depth": "Official abstract and method documentation; generator instructions, structured outputs and scoring path at 30e080f4ecf50912bb0771abb4accbf1e67bb736. Not executed or exhaustively audited.",
    "contribution": "Existing context, derived-variable and relationship generation plus surprise-guided discovery; usable native-method transfer comparator.",
    "limitation": "Model belief shift is not independent truth; current implementation defaults differ from paper experiments.",
    "accessed": "2026-11-15",
    "implementation_url": "https://github.com/allenai/asta-autodiscovery/tree/30e080f4ecf50912bb0771abb4accbf1e67bb736"
  }
]
