Sources we considered, with review status and reasons.
3070 recorded entries · 652 awaiting detailed review · 584 excluded · 672 on hold · 3 publisher withdrawals · 1159 related to curated records.
Suggest a missing contribution or correction.
See omissions and decisions reviewed against public discussion.
Entries without a recorded subject appear only under All subjects. Coverage totals describe all subjects.
1911 matching entries · Page 34 of 39
The current primary disclosure lists drafting/checking proof details and formalization attempts. It does not specifically credit the discretization idea to AI as the older description suggests.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
Jianmu-Q agents derive model mappings and design circuits, but this queue item names the whole system rather than an exact retained mathematical contribution.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
PTagent is an LLM-guided workflow for interpreting existing physics model inputs and producing approved numerical potentials. The tested output is a model-construction tool, with deterministic backends doing the calculations; no AI-origin mathematical result is claimed.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
Current Reddit/HN discussion points to the same benchmark already reviewed. Benchmark performance is not a new theoretical-physics discovery; no duplicate contribution.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-16T07:32:17.125Z
This entry is an aggregate withdrawal notice, not a live proof claim. The original maintainer history identifies a sign error invalidating the shared construction and names all three withdrawn Hodge manuscripts.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
Central AI-origin work in the six-dimensional Tarski lower bound is not identified; the earlier weaker bound was the author’s independent work.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
Human-authored reinforcement-learning algorithm studies model training; no documented AI-origin mathematical contribution in the original claim.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The paper derives a constituent-model hybrid-baryon spectrum, with no accessible disclosure that AI supplied the model or derivation.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The paper introduces an equivariant generative sampler and theoretical guarantees, but does not identify an AI system as originator of those mathematical ideas.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The main text and supplement present a GAN-based neutrino-event surrogate fitted to measurements, without identifying AI as the source of a new mathematical claim.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
IfoScout uses reinforcement learning and differential programming for detector co-design, but the original paper does not attribute its optimization idea or result to an AI research agent.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The accessible abstract reports numerical evidence from stochastic normalizing flows; full text and any AI authorship disclosure were not accessible.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The AI-use statement says ChatGPT discussed arguments and assisted in completing five proofs, but does not identify which central ideas it supplied.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The paper reconstructs Stoic multiple generality using the author’s Natural Term Logic, but the accessible original text does not attribute a central mathematical idea to AI.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
Author says Claude assisted numerical coding and other models checked objections; a human reviewer prompted the sufficiency theorem. No AI-origin central theorem is credited.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
CNet presents a complex-valued ML framework and benchmark results, but the accessible paper does not identify AI-origin research work behind its new layers or Fourier mixer.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The main text describes motif extraction and language editing, but Supplementary Note 10.4 also names AI assistance with proof development. The exact model-origin mathematical work is unresolved.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
AI-use statement expressly limits generative AI to brainstorming, readability, organization and software; authors say it was not an autonomous source of claims or proofs.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The paper analyzes switching between costly LLM oracles, but its abstract and method describe LLMs as objects of study, with no central AI-origin theorem documented.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
Benchmark evaluates logical reasoning on chaotic systems; it does not itself report a new AI-origin scientific or mathematical contribution.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-18T00:00:00.000Z
The acknowledgment of exploring some strategies and calculations does not identify an AI-supplied retained lemma or construction; importance of the sphere theorem does not establish that role.
Screening recorded: 2026-10-16 · Decision reviewed: 2026-10-16T20:44:03.788721Z
The paper proves Jung-constant bounds and a Lean formalization, but the accessible original text does not attribute the mathematical proof to AI.
Screening recorded: 2026-10-17 · Decision reviewed: 2026-10-18T00:00:00.000Z
The AFP entry says Claude initially generated some proofs but people extensively revised them; the retained AI-origin reasoning in the rely-guarantee specification is unclear.
Screening recorded: 2026-10-17 · Decision reviewed: 2026-10-18T00:00:00.000Z
Community guidance and project list is not a single new research contribution.
Screening recorded: 2026-10-17 · Decision reviewed: 2026-10-18T00:00:00.000Z
Nature reports AI-assisted decoding of historical ciphers, but its article redirects to an inaccessible sign-in page, so the exact original artifact and any mathematical contribution cannot be checked.
Screening recorded: 2026-10-17 · Decision reviewed: 2026-10-18T00:00:00.000Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Recovered arXiv metadata screened for mathematics, theoretical computer science and theoretical physics relevance. Not a completed primary-paper review; many are applications, experimental control, language-model evaluations or human work. No publication or proof-validation inference from a metadata match.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-20T07:30:08.090Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Recovered arXiv metadata screened for mathematics, theoretical computer science and theoretical physics relevance. Not a completed primary-paper review; many are applications, experimental control, language-model evaluations or human work. No publication or proof-validation inference from a metadata match.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-20T07:30:08.090Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Recovered PDF describes a fitted knowledge-classification benchmark and AI analysis/code/drafting. A specific central AI theoretical result is not established; reported corpus gains depend on objective and embedding choices.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-23T08:28:30.727Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Promising mathematical-physics lead: reported finiteness for five planar Newtonian bodies at every positive mass, removing the earlier codimension-two exceptional set. General AI use for finding/proving results is disclosed, but contribution-specific allocation of the central valuation/certificate argument and exact external-input/formal semantics need review. Do not infer complete original-proof verification from the paper’s formalization label or numerical interval runs.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Primary v1 states 19,885,160 while recovered metadata states 1,118,430. AI-led optimization and Lean checking are disclosed, but reconcile exact source versions/constants before publishing; the author scopes the gain as an optimization experiment, not the original Linear Hadwiger breakthrough.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-20T07:30:08.090Z
Primary paper and discovery metadata checked. A specific central AI mathematical contribution, exact prior baseline or proof-to-statement scope remains unresolved; Lean use and benchmark performance alone are not promotion criteria.
Screening recorded: 2026-10-19 · Decision reviewed: 2026-10-19T07:23:14.226Z
Read stabilizer-state/finite-field model and AI disclaimer. Model-produced/checking passages do not yet isolate which central Gauss-sum result is AI-origin. Any later record must separate the checked formula subset from the full if-and-only-if and graph-family claims.
Screening recorded: 2026-10-20 · Decision reviewed: 2026-10-20T07:30:08.090Z
Paper credits Claude/Codex for formalization and manuscript preparation; the listing’s AI-assisted proof-search tag does not identify which new construction or proof idea AI supplied. Resolve contribution-specific attribution before inclusion.
Screening recorded: 2026-10-20 · Decision reviewed: 2026-10-20T07:30:08.090Z
Versioned primary paper retrieved. Registry/formalization metadata does not establish contribution-specific central AI origin; isolate the original result, human proof input, prior bound and any changed theorem/checking scope before publication.
Screening recorded: 2026-10-20 · Decision reviewed: 2026-10-20T07:30:08.090Z
The reviewed scope is a proof-agent system, dataset/benchmark, solver or coordination/verification method on supplied targets. No distinct new AI-origin mathematics, TCS theorem or theoretical-physics result is established by that scope; benchmark performance is not promoted as a newly discovered theorem.
Screening recorded: 2026-10-20 · Decision reviewed: 2026-10-20T07:30:08.090Z
Discovery metadata attribution: What AI Found, CC BY 4.0 · VibeMathed contributors, editorial layer CC BY 4.0. AIProofWatch supplies the screening decisions.
Allow optional Google Analytics to help improve AIProofWatch? Privacy details.
