pith. sign in

arxiv: 2606.31273 · v1 · pith:BXC7DRGFnew · submitted 2026-06-30 · 💻 cs.LG

The Calibration Turn in AI-Assisted Research: A Conceptual and Methodological Framework for Evidence-Licensed Claims

Hongmin Li This is my paper

Pith reviewed 2026-07-01 06:25 UTC · model grok-4.3

classification 💻 cs.LG
keywords calibrationevidence-licensed claimsAI-assisted researchepistemic debtclaim-evidence gapassertion rightsscientific claims
0
0 comments X

The pith

Calibration acts as a mechanism for managing scientific assertion rights by licensing claims based on evidence.

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

The paper proposes a framework for AI-assisted research that models the process through five operators: hypothesis generation, model-mediated consequence derivation, external validation, belief update, and claim calibration. It establishes that calibration ensures claims are only made when licensed by evidence, distinguishing this from other forms of semantics. This matters because as AI systems automate more of research, the risk of unlicensed claims grows, and the framework provides principles to evaluate reliability as a closed loop that only outputs evidence-licensed claims.

Core claim

The central claim is that calibration is not merely cautious wording but a mechanism for managing scientific assertion rights: evidence licenses some forms of speech and withholds others. The paper distinguishes linguistic, consequence-based, interventional, and evidence-licensed semantics; defines the claim-evidence gap and epistemic debt; and treats minimal structural reconstruction across heterogeneous outputs as an upward form of claim calibration. The resulting principles are: no claim without license, validation does not determine claim level, and automation amplifies the need for calibration.

What carries the argument

The five operators of AI-assisted research, with claim calibration as the operator that enforces evidence-licensed outputs by managing assertion rights.

Load-bearing premise

That AI-assisted research can be adequately represented by the five operators allowing a coherent distinction between licensed and unlicensed claims.

What would settle it

A counterexample where an AI system produces consistently accurate scientific claims without using any form of claim calibration would challenge the framework's necessity.

Figures

Figures reproduced from arXiv: 2606.31273 by Hongmin Li.

Figure 1
Figure 1. Figure 1: The calibrated AI science loop. Hypothesis generation G proposes candidates from knowledge, accumulated data Dt, and a research question. Modeling MD turns a hypothesis into domain-appropriate testable consequences. Interventional prediction Ph(Y | do(a), x) is one important special case; other domains may derive retrodictive traces, diagnostic features, proof obligations, classification constraints, or me… view at source ↗
Figure 2
Figure 2. Figure 2: Synthetic licensed-utility diagnostic across AISim-Cal routes. All quantities are dimensionless synthetic quantities and not empirical forecasts. Lines show means over the 128 Monte Carlo draws in the illustrative parameterization; propagated standard-error bands are small in this run and are provided in the companion outputs. The curves include goal clarity and route-domain applicability factors; they are… view at source ↗
Figure 3
Figure 3. Figure 3: Calibration ablation in AISim-Cal. All quantities are dimensionless synthetic quantities and not empirical forecasts. The no-calibration, imperfect-calibration, and perfect-calibration conditions show how the synthetic overclaim gap changes when the same route dynamics are passed through different claim-calibration regimes. The calibration ablation is the most direct synthetic illustration of the manuscrip… view at source ↗
Figure 4
Figure 4. Figure 4: AISim-Cal route–domain–scenario score landscape. Each panel corresponds to one scenario and calibration mode; each cell reports the synthetic licensed-utility diagnostic for one route in one domain. The values are dimensionless scores under the chosen synthetic parameterization, including goal clarity and route￾domain applicability Ar,D; they are not empirical forecasts. Hatched rows show low-goal-clarity … view at source ↗
read the original abstract

AI-assisted research has entered a stage in which the central question is not only whether systems can generate hypotheses, run experiments, or produce manuscripts, but whether their scientific claims are calibrated to the evidence that supports them. This Perspective-style paper develops a conceptual and methodological framework for evidence-licensed claims in AI-assisted research. Motivated by representative routes including specialized scientific foundation models, LLM research assistants, multi-agent co-scientists, AI Scientist pipelines, mathematical discovery agents, and self-driving laboratories, it represents AI-assisted research as five operators: hypothesis generation, model-mediated consequence derivation, external validation, belief update, and claim calibration. The central claim is that calibration is not merely cautious wording but a mechanism for managing scientific assertion rights: evidence licenses some forms of speech and withholds others. The paper distinguishes linguistic, consequence-based, interventional, and evidence-licensed semantics; defines the claim-evidence gap and epistemic debt; and treats minimal structural reconstruction across heterogeneous outputs as an upward form of claim calibration. AISim-Cal is included as an illustrative synthetic dynamics exercise, not as an empirical forecast or benchmark. The resulting principles are: no claim without license, validation does not determine claim level, and automation amplifies the need for calibration. Reliable AI-assisted research is therefore evaluated as a loop that generates hypotheses, derives testable consequences, accepts independent adjudication, updates beliefs, and outputs only evidence-licensed claims.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes a conceptual and methodological framework for evidence-licensed claims in AI-assisted research. It models such research via five operators (hypothesis generation, model-mediated consequence derivation, external validation, belief update, and claim calibration) drawn from routes including foundation models, LLM assistants, multi-agent systems, AI Scientist pipelines, mathematical agents, and self-driving labs. The central claim is that calibration functions as a mechanism for managing scientific assertion rights, licensing some claims while withholding others, with supporting distinctions among linguistic, consequence-based, interventional, and evidence-licensed semantics, plus definitions of the claim-evidence gap and epistemic debt. An illustrative synthetic exercise (AISim-Cal) is included, and three principles are derived: no claim without license, validation does not determine claim level, and automation amplifies the need for calibration.

Significance. If the distinctions and operators hold, the framework offers a structured vocabulary for assessing when AI-generated outputs qualify as scientifically assertable, potentially informing evaluation criteria for automated research systems and emphasizing calibration as an active process rather than mere hedging. The definitional approach and illustrative exercise provide a starting point for formalizing epistemic norms in this domain.

major comments (2)
  1. [five operators section] The section defining the five operators: the claim that these operators represent AI-assisted research routes and enable a coherent distinction between evidence-licensed and unlicensed claims rests on an unelaborated assumption of adequacy and coverage; no explicit mapping is provided to the heterogeneous examples (e.g., self-driving laboratories or mathematical discovery agents), which is load-bearing for the central assertion-rights claim.
  2. [central claim and definitions] The definitional structure around calibration and licensing: calibration is introduced in terms of evidence licensing, which is then used to demarcate valid claims, creating a self-referential loop with no independent external benchmark or falsifiable test cited; this directly affects whether the framework can ground the distinction between licensed and unlicensed speech.
minor comments (2)
  1. [abstract and introduction] The abstract and introduction refer to 'Perspective-style paper' and 'illustrative synthetic dynamics exercise' without clarifying how these elements integrate with the methodological framework or whether AISim-Cal is intended to demonstrate any operator in action.
  2. [definitions] The invented terms 'claim-evidence gap' and 'epistemic debt' are introduced without comparison to existing literature on similar concepts (e.g., in philosophy of science or metrology), which would aid readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which identify opportunities to strengthen the presentation of the framework. We address each major comment below, indicating revisions where appropriate.

read point-by-point responses

  1. Referee: [five operators section] The section defining the five operators: the claim that these operators represent AI-assisted research routes and enable a coherent distinction between evidence-licensed and unlicensed claims rests on an unelaborated assumption of adequacy and coverage; no explicit mapping is provided to the heterogeneous examples (e.g., self-driving laboratories or mathematical discovery agents), which is load-bearing for the central assertion-rights claim.

    Authors: We agree that the absence of explicit mappings weakens the load-bearing claim about coverage across routes. In the revised manuscript we will insert a new subsection (following the operator definitions) containing a mapping table. Each row will link one operator to concrete instantiations in the listed routes—for example, external validation in self-driving laboratories occurs via physical assay results, while in mathematical discovery agents it occurs via formal proof checkers. This addition will make the adequacy assumption explicit and support the assertion-rights distinction without altering the core framework. revision: yes

  2. Referee: [central claim and definitions] The definitional structure around calibration and licensing: calibration is introduced in terms of evidence licensing, which is then used to demarcate valid claims, creating a self-referential loop with no independent external benchmark or falsifiable test cited; this directly affects whether the framework can ground the distinction between licensed and unlicensed speech.

    Authors: The concern about circularity is well-taken for a purely definitional treatment. The manuscript already positions external validation as the independent input that feeds the claim-calibration operator; we will revise the definitions section to foreground this sequencing and to state explicitly that licensing is an output of the full operator loop rather than a premise. Because the paper is a conceptual framework rather than an empirical test, no falsifiable benchmark is supplied; we will add a short paragraph clarifying this scope limitation while preserving the internal grounding via the operators. revision: partial

Circularity Check

0 steps flagged

No significant circularity in definitional conceptual framework

full rationale

The paper is a Perspective-style conceptual framework that explicitly defines calibration as a mechanism for evidence-licensed claims and lists resulting principles. No derivation chain, equations, fitted parameters called predictions, or self-citations are present in the provided text. The five operators are introduced as a representation of AI-assisted routes, not derived from or reducing to the central claim. The framework is presented as definitional and illustrative (e.g., AISim-Cal is synthetic and not a benchmark), with no load-bearing step that reduces by construction to its inputs. This matches the default expectation for non-circular conceptual papers.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 2 invented entities

The paper is a perspective that introduces new conceptual distinctions without empirical data or formal derivation. It relies on domain assumptions about how AI research operates and postulates new entities such as epistemic debt without independent evidence.

axioms (2)
  • domain assumption AI-assisted research can be represented as the five operators: hypothesis generation, model-mediated consequence derivation, external validation, belief update, and claim calibration.
    Stated in the abstract as the representation used to develop the framework.
  • domain assumption Evidence can function as a license that permits or withholds specific forms of scientific speech.
    Central to the claim that calibration manages assertion rights.
invented entities (2)
  • claim-evidence gap no independent evidence
    purpose: To quantify or describe the mismatch between asserted claims and supporting evidence.
    Introduced as a defined concept in the framework.
  • epistemic debt no independent evidence
    purpose: To capture accumulated unsupported claims or uncalibrated assertions in AI-assisted outputs.
    Introduced as a defined concept in the framework.

pith-pipeline@v0.9.1-grok · 5780 in / 1524 out tokens · 31472 ms · 2026-07-01T06:25:54.293936+00:00 · methodology

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

68 extracted references · 9 canonical work pages · 4 internal anchors

  1. [1]

    doi:10.1787/a8d820bd-en , url =

    Artificial Intelligence in Science: Challenges, Opportunities and the Future of Research , year =. doi:10.1787/a8d820bd-en , url =

    work page doi:10.1787/a8d820bd-en

  2. [2]

    and Zenil, Hector , title =

    King, Ross D. and Zenil, Hector , title =. Artificial Intelligence in Science: Challenges, Opportunities and the Future of Research , publisher =. 2023 , doi =

    2023

  3. [3]

    and Bambrick, Joshua and Bodenstein, Sebastian W

    Abramson, Josh and Adler, Jonas and Dunger, Jack and Evans, Richard and Green, Tim and Pritzel, Alexander and Ronneberger, Olaf and Willmore, Lindsay and Ballard, Andrew J. and Bambrick, Joshua and Bodenstein, Sebastian W. and Evans, David A. and Hung, Chia-Chun and O'Neill, Michael and Reiman, David and Tunyasuvunakool, Kathryn and Wu, Zachary and Zidek,...

    2024

  4. [4]

    2026 , howpublished =

    2026

  5. [5]

    Nucleic Acids Research , volume =

    Varadi, Mihaly and Bertoni, Damian and Magana, Patricia and Paramval, Urmila and Pidruchna, Ivanna and Radhakrishnan, Mukund and Tsenkov, Maxim and Nair, Sreenath and Mirdita, Milot and Yeo, Josh and Kovalevskiy, Oleg and Tunyasuvunakool, Kathryn and Laydon, Alexander and Zidek, Augustin and Green, Tim and Jumper, John and Birney, Ewan and Steinegger, Mar...

    2024

  6. [6]

    and Aykol, Muratahan and Cheon, Gowoon and Cubuk, Ekin Dogus and others , title =

    Merchant, Amil and Batzner, Simon and Schoenholz, Samuel S. and Aykol, Muratahan and Cheon, Gowoon and Cubuk, Ekin Dogus and others , title =. Nature , volume =. 2023 , doi =

    2023

  7. [7]

    T., Lupsasca, A., Sawhney, M., Scherrer, R., Sellke, M., Spears, B

    Bubeck, Sebastien and Coester, Christian and Eldan, Ronen and Gowers, Timothy and Lee, Yin Tat and Lupsasca, Alexandru and Sawhney, Mehtaab and Scherrer, Robert and Sellke, Mark and Spears, Brian K. and Unutmaz, Derya and Weil, Kevin and Yin, Steven and Zhivotovskiy, Nikita , title =. arXiv preprint arXiv:2511.16072 , year =. doi:10.48550/arXiv.2511.16072 , url =

    work page doi:10.48550/arxiv.2511.16072

  8. [8]

    2025 , howpublished =

    Early Experiments in Accelerating Science with. 2025 , howpublished =

    2025

  9. [9]

    Nature , year =

    Gottweis, Juraj and Weng, Wei-Hung and Daryin, Andrey and others , title =. Nature , year =. doi:10.1038/s41586-026-10644-y , url =

    work page doi:10.1038/s41586-026-10644-y

  10. [10]

    Ghareeb, A. E. and Chang, B. and Mitchener, L. and others , title =. Nature , year =. doi:10.1038/s41586-026-10652-y , url =

    work page doi:10.1038/s41586-026-10652-y

  11. [11]

    Nature , volume =

    Lu, Chris and Lu, Cong and Lange, Robert Tjarko and Yamada, Yutaro and Hu, Shengran and Foerster, Jakob and Ha, David and Clune, Jeff , title =. Nature , volume =. 2026 , doi =

    2026

  12. [12]

    Novikov, Alexander and Vu, Ngan and Eisenberger, Marvin and Dupont, Emilien and Huang, Po-Sen and Wagner, Adam Zsolt and Shirobokov, Sergey and Kozlovskii, Borislav and Ruiz, Francisco J. R. and Mehrabian, Abbas and Kumar, M. Pawan and See, Abigail and Chaudhuri, Swarat and Holland, George and Davies, Alex and Nowozin, Sebastian and Kohli, Pushmeet and Ba...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2506.13131

  13. [13]

    Nature , volume =

    Hubert, Thomas and Mehta, Rupesh and Sartran, Laurent and others , title =. Nature , volume =. 2026 , doi =

    2026

  14. [14]

    and MacKnight, Robert and Kline, Ben and Gomes, Gabe , title =

    Boiko, Daniil A. and MacKnight, Robert and Kline, Ben and Gomes, Gabe , title =. Nature , volume =. 2023 , doi =

    2023

  15. [15]

    and Baird, Sterling G

    Tom, Gary and Schmid, Stefan P. and Baird, Sterling G. and Cao, Yang and Darvish, Kourosh and Hao, Han and Lo, Stanley and Pablo-Garcia, Sergio and Rajaonson, Ella M. and Skreta, Marta and Yoshikawa, Naruki and Corapi, Samantha and Akkoc, Gun Deniz and Strieth-Kalthoff, Felix and Seifrid, Martin and Aspuru-Guzik, Alan , title =. Chemical Reviews , volume ...

    2024

  16. [16]

    and Rendy, B

    Szymanski, Nathan J. and Rendy, B. and Fei, Y. and others , title =. Nature , volume =. 2023 , doi =

    2023

  17. [17]

    and Rendy, Bernardus and Fei, Yuxing and Kumar, Rishi E

    Szymanski, Nathan J. and Rendy, Bernardus and Fei, Yuxing and Kumar, Rishi E. and He, Tanjin and Milsted, David and McDermott, Matthew J. and Gallant, Max and Cubuk, Ekin Dogus and Merchant, Amil and Kim, Haegyeom and Jain, Anubhav and Bartel, Christopher J. and Persson, Kristin and Zeng, Yan and Ceder, Gerbrand , title =. Nature , volume =. 2026 , doi =

    2026

  18. [18]

    Chemical & Engineering News , year =

    Chawla, Dalmeet Singh , title =. Chemical & Engineering News , year =

  19. [19]

    Popular Science Monthly , volume =

    Peirce, Charles Sanders , title =. Popular Science Monthly , volume =

  20. [20]

    , title =

    Popper, Karl R. , title =

  21. [21]

    Criticism and the Growth of Knowledge , editor =

    Lakatos, Imre , title =. Criticism and the Growth of Knowledge , editor =

  22. [22]

    Hacking, Ian , title =

  23. [23]

    , title =

    Mayo, Deborah G. , title =

  24. [24]

    Toulmin, Stephen Edelston , title =

  25. [25]

    Winsberg, Eric , title =

  26. [26]

    Pearl, Judea , title =

  27. [27]

    and Oxman, Andrew D

    Guyatt, Gordon H. and Oxman, Andrew D. and Vist, Gunn E. and Kunz, Regina and Falck-Ytter, Yngve and Alonso-Coello, Pablo and Schunemann, Holger J. , title =. BMJ , volume =. 2008 , doi =

    2008

  28. [28]

    Self-Revising Discovery Systems for Science: A Categorical Framework for Agentic Artificial Intelligence

    Wang, Fiona Y. and Buehler, Markus J. , title =. arXiv preprint arXiv:2606.01444 , year =. doi:10.48550/arXiv.2606.01444 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2606.01444

  29. [29]

    Automatica , volume =

    Rissanen, Jorma , title =. Automatica , volume =. 1978 , doi =

    1978

  30. [30]

    Minimum Description Length Revisited , journal =

    Gr. Minimum Description Length Revisited , journal =. 2019 , doi =

    2019

  31. [31]

    The Journal of Philosophy , volume =

    Friedman, Michael , title =. The Journal of Philosophy , volume =. 1974 , doi =

    1974

  32. [32]

    Philosophy of Science , volume =

    Kitcher, Philip , title =. Philosophy of Science , volume =. 1981 , doi =

    1981

  33. [33]

    , title =

    Thurston, William P. , title =. Bulletin of the American Mathematical Society , volume =. 1994 , doi =

    1994

  34. [34]

    Synthese , volume =

    Avigad, Jeremy , title =. Synthese , volume =. 2006 , doi =

    2006

  35. [35]

    J., Bambrick, J., Bodenstein, S

    Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J., Bambrick, J., Bodenstein, S. W., Evans, D. A., Hung, C.-C., O'Neill, M., Reiman, D., Tunyasuvunakool, K., Wu, Z., Zidek, A., et al. (2024). Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature , 630:493--500

    2024

  36. [36]

    Avigad, J. (2006). Mathematical method and proof. Synthese , 153:105--159

    2006

  37. [37]

    A., MacKnight, R., Kline, B., and Gomes, G

    Boiko, D. A., MacKnight, R., Kline, B., and Gomes, G. (2023). Autonomous chemical research with large language models. Nature , 624:570--578

    2023

  38. [38]

    T., Lupsasca, A., Sawhney, M., Scherrer, R., Sellke, M., Spears, B

    Bubeck, S., Coester, C., Eldan, R., Gowers, T., Lee, Y. T., Lupsasca, A., Sawhney, M., Scherrer, R., Sellke, M., Spears, B. K., Unutmaz, D., Weil, K., Yin, S., and Zhivotovskiy, N. (2025). Early science acceleration experiments with GPT-5 . arXiv preprint arXiv:2511.16072

    work page arXiv 2025

  39. [39]

    Chawla, D. S. (2026). Nature robot chemist paper corrected, but some questions remain unanswered . Chemical & Engineering News . Chemical & Engineering News report, accessed 2026-06-16

    2026

  40. [40]

    Friedman, M. (1974). Explanation and scientific understanding. The Journal of Philosophy , 71(1):5--19

    1974

  41. [41]

    E., Chang, B., Mitchener, L., et al

    Ghareeb, A. E., Chang, B., Mitchener, L., et al. (2026). A multi-agent system for automating scientific discovery. Nature . Advance online publication, published 2026-05-19

    2026

  42. [42]

    AlphaFold protein structure database

    Google DeepMind and EMBL-EBI (2026). AlphaFold protein structure database. https://alphafold.ebi.ac.uk/. Accessed 2026-06-16

    2026

  43. [43]

    Gottweis, J., Weng, W.-H., Daryin, A., et al. (2026). Accelerating scientific discovery with Co-Scientist . Nature . Advance online publication, published 2026-05-19

    2026

  44. [44]

    and Roos, T

    Gr \"u nwald, P. and Roos, T. (2019). Minimum description length revisited. International Journal of Mathematics for Industry , 11(1):1930001

    2019

  45. [45]

    H., Oxman, A

    Guyatt, G. H., Oxman, A. D., Vist, G. E., Kunz, R., Falck-Ytter, Y., Alonso-Coello, P., and Schunemann, H. J. (2008). GRADE : An emerging consensus on rating quality of evidence and strength of recommendations. BMJ , 336(7650):924--926

    2008

  46. [46]

    Hacking, I. (1983). Representing and Intervening: Introductory Topics in the Philosophy of Natural Science . Cambridge University Press, Cambridge

    1983

  47. [47]

    Hubert, T., Mehta, R., Sartran, L., et al. (2026). Olympiad-level formal mathematical reasoning with reinforcement learning. Nature , 651:607--613

    2026

  48. [48]

    King, R. D. and Zenil, H. (2023). A framework for evaluating the AI -driven automation of science. In Artificial Intelligence in Science: Challenges, Opportunities and the Future of Research . OECD Publishing. Accessed 2026-06-16

    2023

  49. [49]

    Kitcher, P. (1981). Explanatory unification. Philosophy of Science , 48(4):507--531

    1981

  50. [50]

    Lakatos, I. (1970). Falsification and the methodology of scientific research programmes. In Lakatos, I. and Musgrave, A., editors, Criticism and the Growth of Knowledge , pages 91--196. Cambridge University Press, Cambridge

    1970

  51. [51]

    T., Yamada, Y., Hu, S., Foerster, J., Ha, D., and Clune, J

    Lu, C., Lu, C., Lange, R. T., Yamada, Y., Hu, S., Foerster, J., Ha, D., and Clune, J. (2026). Towards end-to-end automation of AI research. Nature , 651:914--919

    2026

  52. [52]

    Mayo, D. G. (1996). Error and the Growth of Experimental Knowledge . University of Chicago Press, Chicago

    1996

  53. [53]

    S., Aykol, M., Cheon, G., Cubuk, E

    Merchant, A., Batzner, S., Schoenholz, S. S., Aykol, M., Cheon, G., Cubuk, E. D., et al. (2023). Scaling deep learning for materials discovery. Nature , 624:80--85

    2023

  54. [54]

    AlphaEvolve: A coding agent for scientific and algorithmic discovery

    Novikov, A., Vu, N., Eisenberger, M., Dupont, E., Huang, P.-S., Wagner, A. Z., Shirobokov, S., Kozlovskii, B., Ruiz, F. J. R., Mehrabian, A., Kumar, M. P., See, A., Chaudhuri, S., Holland, G., Davies, A., Nowozin, S., Kohli, P., and Balog, M. (2025). AlphaEvolve : A coding agent for scientific and algorithmic discovery. arXiv preprint arXiv:2506.13131 . T...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  55. [55]

    Artificial intelligence in science: Challenges, opportunities and the future of research

    OECD (2023). Artificial intelligence in science: Challenges, opportunities and the future of research. See the chapter ``A framework for evaluating the AI-driven automation of science''

    2023

  56. [56]

    Early experiments in accelerating science with GPT-5

    OpenAI (2025). Early experiments in accelerating science with GPT-5 . https://openai.com/index/accelerating-science-gpt-5/. Official web report, accessed 2026-06-16

    2025

  57. [57]

    Pearl, J. (2009). Causality: Models, Reasoning, and Inference . Cambridge University Press, Cambridge, 2 edition

    2009

  58. [58]

    Peirce, C. S. (1878). How to make our ideas clear. Popular Science Monthly , 12:286--302

  59. [59]

    Popper, K. R. (1959). The Logic of Scientific Discovery . Hutchinson, London

    1959

  60. [60]

    Rissanen, J. (1978). Modeling by shortest data description. Automatica , 14(5):465--471

    1978

  61. [61]

    J., Rendy, B., Fei, Y., et al

    Szymanski, N. J., Rendy, B., Fei, Y., et al. (2023). An autonomous laboratory for the accelerated synthesis of inorganic materials. Nature , 624:86--91

    2023

  62. [62]

    J., Rendy, B., Fei, Y., Kumar, R

    Szymanski, N. J., Rendy, B., Fei, Y., Kumar, R. E., He, T., Milsted, D., McDermott, M. J., Gallant, M., Cubuk, E. D., Merchant, A., Kim, H., Jain, A., Bartel, C. J., Persson, K., Zeng, Y., and Ceder, G. (2026). Author correction: An autonomous laboratory for the accelerated synthesis of inorganic materials. Nature , 650:E1--E1

    2026

  63. [63]

    Thurston, W. P. (1994). On proof and progress in mathematics. Bulletin of the American Mathematical Society , 30(2):161--178

    1994

  64. [64]

    P., Baird, S

    Tom, G., Schmid, S. P., Baird, S. G., Cao, Y., Darvish, K., Hao, H., Lo, S., Pablo-Garcia, S., Rajaonson, E. M., Skreta, M., Yoshikawa, N., Corapi, S., Akkoc, G. D., Strieth-Kalthoff, F., Seifrid, M., and Aspuru-Guzik, A. (2024). Self-driving laboratories for chemistry and materials science. Chemical Reviews , 124:9633--9732

    2024

  65. [65]

    Toulmin, S. E. (1958). The Uses of Argument . Cambridge University Press, Cambridge

    1958

  66. [66]

    Varadi, M., Bertoni, D., Magana, P., Paramval, U., Pidruchna, I., Radhakrishnan, M., Tsenkov, M., Nair, S., Mirdita, M., Yeo, J., Kovalevskiy, O., Tunyasuvunakool, K., Laydon, A., Zidek, A., Green, T., Jumper, J., Birney, E., Steinegger, M., Hassabis, D., and Velankar, S. (2024). AlphaFold protein structure database in 2024: Providing structure coverage f...

    2024

  67. [67]

    Wang, F. Y. and Buehler, M. J. (2026). Self-Revising Discovery Systems for Science: A Categorical Framework for Agentic Artificial Intelligence . arXiv preprint arXiv:2606.01444

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  68. [68]

    Winsberg, E. (2010). Science in the Age of Computer Simulation . University of Chicago Press, Chicago

    2010