Pith. sign in

REVIEW 5 major objections 5 minor 127 references

AI-for-physics is running the history of discovery in reverse

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 01:15 UTC pith:ZMZ77RXB

load-bearing objection A provocative, self-aware Perspective that reframes AI-for-physics as running history backward; the empirical trend is under-built, but the Reverse ITP concept makes it worth a serious look. the 5 major comments →

arxiv 2607.27794 v1 pith:ZMZ77RXB submitted 2026-07-30 physics.hist-ph cs.AIphysics.pop-ph

Can AI Follow In Einstein's Footsteps?

Michael Shalyt , Nathan Regev , Marin Solja\v{c}i\'c , Ido Kaminer This is my paper
classification physics.hist-ph cs.AIphysics.pop-ph
keywords AI for physicsscientific discoveryparadigm shiftabductive reasoningsymbolic regressionblack-box predictiontheory buildingphilosophy of physics
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that the most visible AI contributions to physics are recapitulating the history of physics backwards. Human physics moved from pattern prediction to phenomenological laws to principle-based universal theories, while AI-for-physics has moved from explicit equation discovery (symbolic regression) to powerful but opaque neural predictors. As a result, current AI systems excel at induction and deduction but cannot perform the abductive, principle-guided reasoning behind general relativity or the Standard Model. The paper classifies discoveries into three categories — new solutions, new equations, new mathematical frameworks — and claims no AI has yet made a Category C discovery. Its practical thesis is that better prediction alone will not yield paradigm-level theories; AI needs to propose principles, pose questions, and search over provisional axioms.

Core claim

On the authors' own terms, the central claim is that the center of gravity of AI for physics discovery has shifted from discovering explicit symbolic laws to building black-box predictors, reversing the epistemic order of human physics. Consequently, while AI has produced Category A discoveries (predictions, devices) and Category B discoveries (equations, laws), no AI system has produced a Category C discovery — the successful use of a new principle or the invention of a novel mathematical framework. The paper does not claim this is impossible; it claims the field is currently optimized away from it. It argues that the missing skill is not creativity but scientific taste: choosing which nove

What carries the argument

The central machinery is a three-tier classification of physics discoveries (Category A: new solutions or capabilities; Category B: new equations or laws; Category C: new mathematical frameworks or principles), together with a historical reversal thesis that maps human physics' progression — pattern prediction, phenomenology, principle-based theory — onto AI's chronological trajectory in reverse. The classification locates the gap: AI has achieved A and B, never C. The concrete mechanism proposed to close the gap is a 'Reverse ITP': a formal system that organizes abductive search over provisional axioms rather than proving consequences from fixed axioms, using contradictions as a loss signal

Load-bearing premise

The claim depends on the sample of 'most visible and influential' AI-for-science successes being representative of the field's center of gravity; if that sample is biased, the reverse trajectory may be a selection artifact rather than a real trend.

What would settle it

A published, replicated AI-generated physics result that introduces a new mathematical framework or symmetry principle later validated by experiment would falsify the claim that no AI has made a Category C discovery.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the reversal thesis holds, continued investment in black-box prediction alone will not produce paradigm-level theories like quantum gravity.
  • AI systems that can propose and test provisional axioms (reverse ITPs) would enable automated generation of falsifiable theory candidates.
  • A '1911 cutoff' test — withholding general relativity from training data and seeing whether AI rediscovers it — becomes a meaningful benchmark for principle-level discovery.
  • If AI can discover simple theories in artificial worlds but not real-world ones, the bottleneck may be physics itself, not AI architecture.
  • The three-category taxonomy gives the field a concrete target: explicitly aiming for Category C discovery as a stated research goal.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The reversal thesis may be partly a selection artifact: a broader sample that includes recent symbolic-regression and automated-theorem-proving successes from the same period could weaken or reverse the apparent trajectory.
  • A testable extension is corpus-level measurement of AI-for-physics outputs over time — coding each contribution by category and by whether it outputs explicit equations — to see whether the trend is robust or depends on which successes are counted.
  • The 'Reverse ITP' concept suggests a new kind of AI benchmark: not solving problems, but generating provisional axioms whose falsifiable consequences are novel and testable.
  • The paper's own caveat about non-human-interpretable mathematics implies that human explainability may become a bottleneck to paradigm shifts rather than a necessity; if AI develops powerful non-human-interpretable frameworks, the field's target might shift to detecting and validating discoveries through compression measures.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. This Perspective argues that AI contributions to physics discovery are following the historical trajectory of human physics in reverse epistemic order: early AI milestones extracted explicit symbolic laws (BACON, Eureqa, SINDy, AI Feynman), while the most prominent recent systems (AlphaFold, GraphCast, GNoME) are high-accuracy black-box predictors. The paper introduces an A/B/C taxonomy of physics discoveries (new solutions/capabilities, new equations/laws, new mathematical frameworks/principle-based theories) and asserts that no AI has yet made a Category C discovery. It concludes that continued progress in prediction alone will not yield paradigm-level theories such as quantum gravity, and proposes a research agenda to give AI systems skills for posing questions, inventing principles, conducting abductive searches over provisional axioms, and using symbolic-computing theory-building sandboxes, including a 'Reverse ITP'. The paper is framed as a Perspective and includes explicit caveats that the historical sketch is approximate and the reversal is not a universal law.

Significance. The paper addresses a timely and important question: whether current AI-for-science paradigms, centered on predictive accuracy, can lead to fundamental theoretical breakthroughs in physics. If the central claim is correct, it implies a need to reallocate research effort toward principle discovery, symbolic reasoning, falsifiable theory generation, and formalized abduction. The paper's strengths are its clear historical framing, a useful (if underspecified) taxonomy, a concrete falsifiable proposal (e.g., the '1911 cutoff' test and artificial-world tasks), and several constructive technical suggestions (theory-building sandboxes, symmetry abduction). However, the central empirical assertions rest on a hand-selected sample and non-operational categories; these need to be tightened before the paper's conclusions can be fully endorsed.

major comments (5)
  1. [§2, Fig. 1] The reverse-trajectory claim is supported by a small set of 'most visible' AI successes (AlphaFold, GraphCast, GNoME; refs 21–24, 62) rather than a systematic survey. The paper's own references include recent symbolic-regression and analytical theory-building work (refs 18, 19, 44, 45, 106), so the assertion that black-box systems 'increasingly overshadow' these efforts is asserted, not demonstrated. The paper correctly labels the sketch 'approximate' and the reversal 'not a universal law', but these qualifications do not replace evidence. A bibliometric or corpus-based analysis, or at least a falsifiable sampling protocol, is needed to establish that the trend is real and representative. This is load-bearing because the central prediction depends on the trend.
  2. [§4, Table 1; §5] The claim 'no AI has made a discovery of Category C' is not operationally testable as stated. Category C is defined only by examples ('successful use of a new principle or the invention of a novel mathematical framework'), with no criteria for 'successful use', 'new principle', or 'novel framework'. Because the table is constructed with all current AI outputs placed in Categories A and B, the absence of Category C is partly a consequence of the classification. The paper itself says the categories are not strict or mutually exclusive, which makes the universal negative even harder to evaluate. In addition, Table 1 lists the Standard Model and Higgs mechanism as human Category B, yet Section 5 cites the Higgs mechanism as an example of principle-guided abductive discovery; this internal inconsistency blurs the taxonomy. The authors should provide operational criteria and a systematic audit
  3. [§6, ref. [114]] A key demonstration for the proposed roadmap (inverting the EFT workflow to search for candidate quantum-gravity theories) is supported by reference [114], which is an unpublished 'in preparation' self-citation. This is not verifiable by readers or reviewers. The authors should either make the work available (e.g., as a preprint or extended supporting information) or describe the method and results in sufficient detail to be evaluated. As it stands, the feasibility of the proposed direction is partly grounded in an inaccessible source.
  4. [§2–§3] The paper's sample mixes AI-for-science successes (AlphaFold for protein folding, GraphCast for weather, GNoME for materials) with AI-for-physics work. These are impressive predictive applications but are not discoveries of physical laws in the sense used in the human trajectory (Kepler, Maxwell, Einstein). Including them as the 'frontier' of AI-for-physics biases the sample toward prediction. The authors should either restrict the claim to physics-specific AI systems or explicitly justify why cross-domain examples are representative of the center of gravity of AI-for-physics.
  5. [§6, 'Is the Bottleneck AI, or Physics Itself?'] The paper acknowledges that the absence of new paradigm-level theories may be due to physics itself (the end of simple, testable revolutions) rather than to AI limitations, but it leaves this possibility largely unintegrated. If the bottleneck is the intrinsic complexity or experimental untestability of remaining theories, then the reverse trajectory does not explain or predict the absence of AI Category C discoveries. The authors should state what evidence would distinguish the 'AI bottleneck' from the 'physics bottleneck' hypotheses—their artificial-world and '1911 cutoff' proposals are a good start—and explicitly condition the central claim on the outcome of such tests.
minor comments (5)
  1. [Fig. 1] The figure would benefit from labeled axes and a legend distinguishing the human and AI timelines. The caption refers to 'arrows summarizing example contributions' but the arrows are not individually identified.
  2. [§4, Historical Examples] The phrase 'a certain ansätze' should be 'a certain ansatz' (ansätze is the plural form).
  3. [§6] The heading 'Propose questions rather than answers' would read more naturally as 'Proposing questions rather than answers.'
  4. [§3, 'Ipcha Mistabra'] The caveat is valuable, but its connection to the surrounding argument could be made explicit: it currently interrupts the flow between the Bunge discussion and the alignment paragraph.
  5. [§2] The phrase 'reverse epistemic order' is used interchangeably with 'reverse order' and 'reverse trajectory'; consider defining the intended meaning of 'epistemic' early to avoid ambiguity.

Circularity Check

0 steps flagged

No significant circularity: the paper's historical and categorical claims are explicitly framed as approximate empirical interpretations, not as derivations from their own definitions.

full rationale

This Perspective does not contain a derivation chain in which a fitted parameter is relabeled as a prediction or in which a target result is assumed by construction. The central claims—that the most visible AI-for-physics successes have moved from symbolic equation discovery toward black-box predictors, and that no AI has yet produced a Category C discovery—are presented as empirical and interpretive, with repeated caveats: the reversal is 'not a universal law' (Sec. 2), the historical sketch is 'necessarily approximate' (Sec. 2), and the A/B/C categories 'do not serve as strict mutually exclusive classifications' (Sec. 4). The empty Category C entry for AI is an observed absence, not a logical consequence of the category definitions: the definitions are given in terms of human examples (calculus, gauge theory, Hilbert space), and the paper explicitly states that AI could in principle occupy Category C. The single self-citation to Ref. [114], an in-preparation work by the corresponding author, is used only as an illustrative example of a symbolic search over EFTs ('this approach illustrates how symbolic infrastructure can support the discovery of new candidate theories'); it is not the load-bearing justification for the paper's main thesis, which rests on the historical trend and the claimed absence of Category C discoveries. Concerns about hand-picked examples or the operational vagueness of 'new principle' are legitimate correctness or falsifiability criticisms, but they are not instances of circularity under the stated criteria. The paper is self-contained as a perspective piece: it argues from external, checkable milestones (AlphaFold, GraphCast, SINDy, Eureqa, BACON, etc.) and does not reduce its conclusions to its own definitions or citations.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 1 invented entities

The central argument depends on a set of interpretive assumptions about the history of physics, the representativeness of selected AI milestones, and the mode of human discovery. None are numerically fitted; the 'free parameter' is the unstated selection rule for examples. The Reverse ITP is an invented entity with no independent evidence.

axioms (5)
  • domain assumption The history of physics can be coarsely ordered as pattern prediction -> phenomenological laws -> principle-based theory, with increasing understanding.
    Section 2 and Fig. 1; the reversal claim depends on this ordering. The paper itself notes ancient astronomers held principle-based views, weakening the ordering as a strict benchmark.
  • domain assumption The 'most visible and influential' AI-for-science successes (AlphaFold, GraphCast, GNoME, Project NeRD) represent the field's center of gravity.
    Section 2; the reverse trajectory is inferred from this selection, which is not based on a systematic survey or quantitative corpus.
  • domain assumption A black-box predictor cannot undergo a Kuhnian crisis and therefore cannot produce a paradigm shift.
    Section 3, citing Bunge's 'A General Black Box Theory' [75]; used to conclude that pure prediction 'will struggle' to produce quantum gravity-level theories.
  • domain assumption Einstein, Dirac, and Higgs used abduction (principle-driven conjecture) rather than induction or deduction, and this mode is necessary for Category C discoveries.
    Section 4; supports the claim that current AIs lack the key skill. Historical scholarship might dispute this reconstruction, and the paper itself cites Maxwell's discarded mechanical scaffolding.
  • domain assumption LLMs are trained to optimize for mainstream consensus (the statistical mode) and smooth away outlier 'weirdness'.
    Section 6; underpins the 'training to the tail' discussion but is asserted without citation in this strong form.
invented entities (1)
  • Reverse ITP no independent evidence
    purpose: A proposed formal system that formalizes abductive search over provisional axioms, allowing AI to propose temporary axioms and derive falsifiable consequences.
    Section 7; not implemented, no specification, and no falsifiable handle outside the paper. It is the paper's central invented construct for enabling Category C discovery.

pith-pipeline@v1.3.0-daily-deepseek · 15434 in / 12201 out tokens · 108410 ms · 2026-08-01T01:15:35.393110+00:00 · methodology

0 comments
read the original abstract

AI is accelerating physics discovery, but perhaps away from Einstein-level theory building. To understand this gap, we must recognize a striking trend: while being very successful, the most visible AI contributions to physics discovery appear to mirror the historical development of physics, but in reverse. Human discovery in physics progressed, in broad strokes, from ancient pattern prediction, through phenomenological laws such as Kepler's, to principle-based universal theories such as relativity and the Standard Model. On the AI side, prominent contributions to physics discovery point in the opposite direction: early milestones emphasized explicit equation-discovery methods, such as symbolic regression, whereas more recent frontier contributions are powerful predictors such as AlphaFold and GraphCast, which can be remarkably accurate yet do not provide clear theoretical understanding. If this trend continues, AI would become extraordinarily good at prediction but may struggle to ever propose its first serious contender to quantum gravity or other paradigm-level theories. We review the current landscape of AI for physics discovery and highlight a critical missing skill: the ability to pose the right questions or invent the right principles to guide the development of new theories and the tests to falsify them. This mode of discovery has driven many of the deepest advances since the 17th century, where symmetry, simplicity, and new mathematical frameworks guided theory construction before experimental tests. Equipping AI systems with such skills could move them from predicting within known frameworks to proposing the next paradigm-level discovery in physics.

Figures

Figures reproduced from arXiv: 2607.27794 by Ido Kaminer, Marin Solja\v{c}i\'c, Michael Shalyt, Nathan Regev.

Figure 1
Figure 1. Figure 1: The trajectories of human vs AI physics discovery – reversal in explainability. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Evolution of scientific theories. A schematic representation of the iterative, upward growth of scientific understanding. Every extension of an established theory provides predictions that are tested against observations, either refuting this theory ex￾tension (tree stump) or upholding it (con￾tinued growth of the trunk). Every exten￾sion of the theory provides further applica￾tions (fruits) and increased … view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

127 extracted references · 10 linked inside Pith

  1. [1]

    Wang, H.et al.Scientific discovery in the age of artificial intelligence.Nature620,47–60 (2023)

    2023

  2. [2]

    & King, R

    Kramer, S., Cerrato, M., Brugger, J., Džeroski, S. & King, R. D. Automated Scientific Dis- covery: From Equation Discovery to Autonomous Discovery Systems.Machine Learning 115,109 (2026)

    2026

  3. [3]

    & Tegmark, M

    Udrescu, S.-M. & Tegmark, M. AI Feynman: A physics-inspired method for symbolic regres- sion.Science Advances6(2020)

    2020

  4. [4]

    Artificial intelligence to win the Nobel Prize and beyond: Creating the engine of scientific discovery.AI magazine37,39–49 (2016)

    Kitano, H. Artificial intelligence to win the Nobel Prize and beyond: Creating the engine of scientific discovery.AI magazine37,39–49 (2016)

    2016

  5. [5]

    Carleo, G.et al.Machine learning and the physical sciences.Reviews of Modern Physics91, 045002 (2019)

    2019

  6. [6]

    & Haken, W

    Appel, K. & Haken, W. Solution of the four color map problem.Scientific American237, 108–121 (1977)

    1977

  7. [7]

    & von Raumer, J.The Lean theorem prover (system description)inInternational Conference on Automated Deduction(2015), 378–388

    De Moura, L., Kong, S., Avigad, J., Van Doorn, F. & von Raumer, J.The Lean theorem prover (system description)inInternational Conference on Automated Deduction(2015), 378–388

    2015

  8. [8]

    On Conjectures of Graffiti.Annals of Discrete Mathematics38,113–118 (1988)

    Fajtlowicz, S. On Conjectures of Graffiti.Annals of Discrete Mathematics38,113–118 (1988)

    1988

  9. [9]

    Raayoni, G.et al.Generating conjectures on fundamental constants with the Ramanujan Ma- chine.Nature590,67–73 (2021)

    2021

  10. [10]

    Elimelech, R.et al.Algorithm-assisted discovery of an intrinsic order among mathematical constants.Proceedings of the National Academy of Sciences121,e2321440121 (2024)

    2024

  11. [11]

    Davies, A.et al.Advancing mathematics by guiding human intuition with AI.Nature600, 70–74 (2021)

    2021

  12. [12]

    Fawzi, A.et al.Discovering faster matrix multiplication algorithms with reinforcement learn- ing.Nature610,47–53 (2022)

    2022

  13. [13]

    & Aarrestad, T

    Belis, V ., Odagiu, P. & Aarrestad, T. K. Machine learning for anomaly detection in particle physics.Reviews in Physics12,100091 (2024)

    2024

  14. [14]

    Langley, P.BACON: A production system that discovers empirical lawsinProceedings of the 5th International Joint Conference on Artificial Intelligence(1977), 344

    1977

  15. [15]

    A., Bradshaw, G

    Langley, P., Simon, H. A., Bradshaw, G. L. & Zytkow, J. M.Scientific Discovery: Computa- tional Explorations of the Creative Process(MIT Press, 1987)

    1987

  16. [16]

    & Lipson, H

    Schmidt, M. & Lipson, H. Distilling Free-Form Natural Laws from Experimental Data.Sci- ence324,81–85 (2009)

    2009

  17. [17]

    L., Proctor, J

    Brunton, S. L., Proctor, J. L. & Kutz, J. N. Discovering governing equations from data by sparse identification of nonlinear dynamical systems.Proceedings of the National Academy of Sciences113,3932–3937 (2016)

    2016

  18. [18]

    Li, Q.et al.Advancing symbolic regression for earth science with a focus on evapotranspira- tion modeling.npj Climate and Atmospheric Science7,275 (2024)

    2024

  19. [19]

    Advances in Neural Information Processing Systems33,17429–17442 (2020)

    Cranmer, M.et al.Discovering Symbolic Models from Deep Learning with Inductive Biases. Advances in Neural Information Processing Systems33,17429–17442 (2020)

    2020

  20. [20]

    The Computational Gauntlet of Human-Like Learning.Proceedings of the 36th AAAI Conference on Artificial Intelligence(2022)

    Langley, P. The Computational Gauntlet of Human-Like Learning.Proceedings of the 36th AAAI Conference on Artificial Intelligence(2022)

    2022

  21. [21]

    W., Evans, R., Jumper, J.,et al.Improved protein structure prediction using poten- tials from deep learning.Nature577,706–710 (2020)

    Senior, A. W., Evans, R., Jumper, J.,et al.Improved protein structure prediction using poten- tials from deep learning.Nature577,706–710 (2020)

    2020

  22. [22]

    Jumper, J.et al.Highly Accurate Protein Structure Prediction with AlphaFold.Nature596, 583–589 (2021)

    2021

  23. [23]

    Abramson, J., Adler, J., Dunger, J.,et al.Accurate structure prediction of biomolecular inter- actions with AlphaFold 3.Nature630,493–500 (2024)

    2024

  24. [24]

    Lam, R.et al.Learning skillful medium-range global weather forecasting.Science382,1416– 1421 (2023)

    2023

  25. [25]

    & Karniadakis, G

    Raissi, M., Perdikaris, P. & Karniadakis, G. E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations.Journal of Computational Physics378,686–707 (2019). 13

    2019

  26. [26]

    & Karniadakis, G

    Raissi, M., Yazdani, A. & Karniadakis, G. E. Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations.Science367,1026–1030 (2020)

    2020

  27. [27]

    Peurifoy, J.et al.Nanophotonic particle simulation and inverse design using artificial neural networks.Science Advances4,eaar4206 (2018)

    2018

  28. [28]

    V ., Das, P

    Pestourie, R., Mroueh, Y ., Nguyen, T. V ., Das, P. & Johnson, S. G. Active learning of deep surrogates for PDEs: application to metasurface design.npj Computational Materials6,1–7 (2020)

    2020

  29. [29]

    Jiang, R. & Willett, R.Embed and emulate: learning to estimate parameters of dynamical systems with uncertainty quantificationinProceedings of the 36th International Conference on Neural Information Processing Systems(Curran Associates Inc., 2022)

    2022

  30. [30]

    A., Ricardez-Sandoval, L

    Elorza Casas, C. A., Ricardez-Sandoval, L. A. & Pulsipher, J. L. A comparison of strategies to embed physics-informed neural networks in nonlinear model predictive control formulations solved via direct transcription.Computers and Chemical Engineering198,109105 (2025)

    2025

  31. [31]

    Wei, J.et al.Chain-of-thought prompting elicits reasoning in large language models.Ad- vances in Neural Information Processing Systems35,24824–24837 (2022)

    2022

  32. [32]

    A., MacKnight, R

    Boiko, D. A., MacKnight, R. & Gomes, G.Emergent autonomous scientific research capa- bilities of large language modelsarXiv:2304.05332 [physics.chem-ph](2023)

    Pith/arXiv arXiv 2023

  33. [33]

    M.et al.Augmenting large language models with chemistry tools.Nature Machine Intelligence6,525–535 (2024)

    Bran, A. M.et al.Augmenting large language models with chemistry tools.Nature Machine Intelligence6,525–535 (2024)

    2024

  34. [34]

    Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language ModelsarXiv:2501.09686 [cs.AI](2025)

    Xu, F.et al. Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language ModelsarXiv:2501.09686 [cs.AI](2025)

    Pith/arXiv arXiv 2025

  35. [35]

    Solving Quantitative Reasoning Problems with Language Modelsin Advances in Neural Information Processing Systems(2022)

    Lewkowycz, A.et al. Solving Quantitative Reasoning Problems with Language Modelsin Advances in Neural Information Processing Systems(2022)

    2022

  36. [36]

    Zheng, K., Han, J. M. & Polu, S.miniF2F: a cross-system benchmark for formal Olympiad- level mathematicsinInternational Conference on Learning Representations(2022)

    2022

  37. [37]

    H., Wu, Y ., Le, Q

    Trinh, T. H., Wu, Y ., Le, Q. V ., He, H. & Luong, T. Solving olympiad geometry without human demonstrations.Nature625,476–482 (2024)

    2024

  38. [38]

    From Euler to AI: Unifying Formulas for Mathematical ConstantsinAdvances in Neural Information Processing Systems38(Curran Associates, Inc., 2025), 123959–124040

    Raz, T.et al. From Euler to AI: Unifying Formulas for Mathematical ConstantsinAdvances in Neural Information Processing Systems38(Curran Associates, Inc., 2025), 123959–124040

    2025

  39. [39]

    & Sutskever, I.Generative Language Modeling for Automated Theorem Proving arXiv:2009.03393 [cs.LG](2020)

    Polu, S. & Sutskever, I.Generative Language Modeling for Automated Theorem Proving arXiv:2009.03393 [cs.LG](2020)

    Pith/arXiv arXiv 2009

  40. [40]

    Reddy, C. K. & Shojaee, P.Towards scientific discovery with generative AI: progress, op- portunities, and challengesinProceedings of the Thirty-Ninth AAAI Conference on Artificial Intelligence and Thirty-Seventh Conference on Innovative Applications of Artificial Intelli- gence and Fifteenth Symposium on Educational Advances in Artificial Intelligence(AAA...

    2025

  41. [41]

    Sothanaphan, N.Resolution of Erdos Problem 728: a writeup of Aristotle’s Lean proofarXiv: 2601.07421 [math.NT](2026)

    arXiv 2026

  42. [42]

    Fel’s Conjecture on Syzygies of Numerical SemigroupsarXiv:2602.03716 [math.CO](2026)

    Chen, E.et al. Fel’s Conjecture on Syzygies of Numerical SemigroupsarXiv:2602.03716 [math.CO](2026)

    arXiv 2026

  43. [43]

    & Tang, Q.An Erdos problem on random subset sums in finite abelian groupsarXiv: 2602.05768 [math.CO](2026)

    Ma, J. & Tang, Q.An Erdos problem on random subset sums in finite abelian groupsarXiv: 2602.05768 [math.CO](2026)

    arXiv 2026

  44. [44]

    & Weil, K.Single-minus gluon tree amplitudes are nonzeroarXiv:2602.12176 [hep-th](2026)

    Guevara, A., Lupsasca, A., Skinner, D., Strominger, A. & Weil, K.Single-minus gluon tree amplitudes are nonzeroarXiv:2602.12176 [hep-th](2026)

    arXiv 2026

  45. [45]

    & Weil, K.Single-minus graviton tree amplitudes are nonzeroarXiv:2603.04330 [hep-th](2026)

    Guevara, A., Lupsasca, A., Skinner, D., Strominger, A. & Weil, K.Single-minus graviton tree amplitudes are nonzeroarXiv:2603.04330 [hep-th](2026)

    arXiv 2026

  46. [46]

    Shen, Y .et al.Deep learning with coherent nanophotonic circuits.Nature Photonics11,441– 446 (2017)

    2017

  47. [47]

    & Sohl-Dickstein, J

    Bahri, Y ., Kadmon, J., Ganguli, S. & Sohl-Dickstein, J. Statistical mechanics of deep learning. Annual Review of Condensed Matter Physics11,501–528 (2020)

    2020

  48. [48]

    Toscano, J., Oommen, V ., Varghese, A.,et al.From PINNs to PIKANs: recent advances in physics-informed machine learning.Machine Learning and Computational Science & Engi- neering1,15 (2025)

    2025

  49. [49]

    KAN: Kolmogorov-Arnold NetworksinInternational Conference on Learning Representations (ICLR)(2025)

    Liu, Z.et al. KAN: Kolmogorov-Arnold NetworksinInternational Conference on Learning Representations (ICLR)(2025). 14

    2025

  50. [50]

    & Karniadakis, G

    Raissi, M., Perdikaris, P., Ahmadi, N. & Karniadakis, G. E. Physics-Informed Neural Net- works and Extensions.arXiv:2408.16806(2024)

    Pith/arXiv arXiv 2024

  51. [51]

    Burger, B.et al.A mobile robotic chemist.Nature583,237–241 (2020)

    2020

  52. [52]

    A., MacKnight, R., Kline, B

    Boiko, D. A., MacKnight, R., Kline, B. & Gomes, G. Autonomous chemical research with large language models.Nature624,570–578 (2023)

    2023

  53. [53]

    & Shih, D

    Karagiorgi, G., Kasieczka, G., Kravitz, S., Nachman, B. & Shih, D. Machine learning in the search for new fundamental physics.Nature Reviews Physics4,399–412 (2022)

    2022

  54. [54]

    & Hey, T

    Thiyagalingam, J., Shankar, M., Fox, G. & Hey, T. Scientific machine learning benchmarks. Nature Reviews Physics4,413–420 (2022)

    2022

  55. [55]

    & Aspuru-Guzik, A

    Krenn, M., Pollice, R., Häse, F., Friederich, P. & Aspuru-Guzik, A. On scientific understand- ing with artificial intelligence.Nature Reviews Physics4,761–769 (2022)

    2022

  56. [56]

    & Beucler, T

    Gentine, P., Eyring, V . & Beucler, T. Machine learning for the physics of climate.Nature Reviews Physics6,249–251 (2024)

    2024

  57. [57]

    Jiao, L.et al.AI meets physics: A comprehensive survey.Artificial Intelligence Review57, 1–70 (2024)

    2024

  58. [58]

    & Chawla, S.A Perspective on Symbolic Machine Learning in Physical Sciences inNeurIPS 2024 Workshop on Machine Learning and the Physical Sciences(2024)

    Makke, N. & Chawla, S.A Perspective on Symbolic Machine Learning in Physical Sciences inNeurIPS 2024 Workshop on Machine Learning and the Physical Sciences(2024)

    2024

  59. [59]

    J., Ha, S., Iten, R., Klopotek, M

    Wetzel, S. J., Ha, S., Iten, R., Klopotek, M. & Liu, Z.Interpretable Machine Learning in Physics: A ReviewarXiv:2503.23616 [physics.comp-ph](2025)

    Pith/arXiv arXiv 2025

  60. [60]

    Ray, S. Generative Metascience: A Review of AI as the Next Scientific Instrument and the Emerging Paradigm of Algorithmic Discovery.MetaScientia: Journal of the History and Phi- losophy of Science1,249–285 (2025)

    2025

  61. [61]

    Newton, I.Philosophiae naturalis principia mathematica(Jussu Societatis Regiae ac Typis Josephi Streater, 1687)

  62. [62]

    Merchant, A.et al.Scaling deep learning for materials discovery.Nature624,80–85 (2023)

    2023

  63. [63]

    Maxwell, J. C. On Physical Lines of Force. Part I.–The Theory of Molecular V ortices applied to Magnetic Phenomena.The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science. 421,161–175 (1861)

  64. [64]

    & Crockett, M

    Messeri, L. & Crockett, M. J. Artificial intelligence and illusions of understanding in scientific research.Nature627,49–58 (2024)

    2024

  65. [65]

    Langley, P.Integrated Systems for Computational Scientific DiscoveryinProceedings of the AAAI Conference on Artificial Intelligence(2024)

    2024

  66. [66]

    Cai, S., Mao, Z., Wang, Z.,et al.Physics-informed neural networks (PINNs) for fluid me- chanics: a review.Acta Mechanica Sinica37,1727–1738 (2021)

    2021

  67. [67]

    E.et al.Physics-informed machine learning.Nature Reviews Physics3,422– 440 (2021)

    Karniadakis, G. E.et al.Physics-informed machine learning.Nature Reviews Physics3,422– 440 (2021)

    2021

  68. [68]

    & Solja ˇci´c, M

    Loh, C., Christensen, T., Dangovski, R., Kim, S. & Solja ˇci´c, M. Surrogate- and invariance- boosted contrastive learning for data-scarce applications in science.Nature Communications 13,4223 (2022)

    2022

  69. [69]

    Lagrangian Neural NetworksarXiv:2003.04630 [cs.LG](2020)

    Cranmer, M.et al. Lagrangian Neural NetworksarXiv:2003.04630 [cs.LG](2020)

    Pith/arXiv arXiv 2003

  70. [70]

    Neural Robot Dynamicsin9th Annual Conference on Robot Learning(2025)

    Xu, J.et al. Neural Robot Dynamicsin9th Annual Conference on Robot Learning(2025)

    2025

  71. [71]

    Learning to Simulate Complex Physics with Graph Networksin Proceedings of the 37th International Conference on Machine Learning119(PMLR, 2020), 8459–8468

    Sanchez-Gonzalez, A.et al. Learning to Simulate Complex Physics with Graph Networksin Proceedings of the 37th International Conference on Machine Learning119(PMLR, 2020), 8459–8468

    2020

  72. [72]

    & Edmunds, M

    Seiradakis, J. & Edmunds, M. Our current knowledge of the Antikythera Mechanism.Nature Astronomy2,35–42 (2018)

    2018

  73. [73]

    Sutton, R.The Bitter Lesson(Mar. 2019). www.incompleteideas.net/IncIdeas/BitterLesson.html

    2019

  74. [74]

    Yousefi, M. & Collins, J.Learning the Bitter Lesson: Empirical Evidence from 20 Years of CVPR ProceedingsinProceedings of the 1st Workshop on NLP for Science (NLP4Science) (Association for Computational Linguistics, 2024), 175–187

    2024

  75. [75]

    A General Black Box Theory.Philosophy of Science30,346–358 (1963)

    Bunge, M. A General Black Box Theory.Philosophy of Science30,346–358 (1963)

    1963

  76. [76]

    S.The Structure of Scientific Revolutions(University of Chicago Press, 1962)

    Kuhn, T. S.The Structure of Scientific Revolutions(University of Chicago Press, 1962)

    1962

  77. [77]

    Emergent Communication at ScaleinInternational Conference on Learning Representations(2022)

    Chaabouni, R.et al. Emergent Communication at ScaleinInternational Conference on Learning Representations(2022). 15

    2022

  78. [78]

    Lee, J., Cho, K. & Kiela, D.Countering Language Drift via Visual GroundinginProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)(Asso- ciation for Computational Linguistics, 2019), 4385–4395

    2019

  79. [79]

    Training Large Language Models to Reason in a Continuous Latent Space arXiv:2412.06769 [cs.CL](2025)

    Hao, S.et al. Training Large Language Models to Reason in a Continuous Latent Space arXiv:2412.06769 [cs.CL](2025)

    Pith/arXiv arXiv 2025

  80. [80]

    Laughlin, R. B. Anomalous Quantum Hall Effect: An Incompressible Quantum Fluid with Fractionally Charged Excitations.Physical Review Letters50,1395–1398 (May 1983)

    1983

Showing first 80 references.