Pith. sign in

REVIEW 2 major objections 4 minor 44 references

Machine Learning is Good for Physics - and Vice Versa

T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper argues that machine learning should extend—not replace—the two pillars of fundamental physics: controlled statistical inference and generalizing theory interpretation.

desk verdict A solid programmatic essay: the four-way ML taxonomy and the defense of statistical and theory standards are useful; the agentic-hypothesis boundary needs work. read the letter →

arxiv 2608.05812 v1 pith:6UINPQ6X submitted 2026-08-06 hep-ph

Michael Krämer , Tilman Plehn This is my paper · ORCID
classification hep-ph
keywords machinelearningfundamentalphysicsparticlequantumfieldtheorystatisticalinferenceanomalydetectionrepresentationsimulation-based
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The essay argues that machine learning should be integrated into fundamental physics as an extension of its toolset, not as a replacement for theory or statistical method. Its central assertion is that ML 'do[es] not replace the role of theoretical structures, but extend[s] the set of tools through which they can be connected to data,' and that two pillars—controlled statistical inference and generalizing theory interpretation—must not be weakened. The paper surveys three ways ML enters particle physics—enhancing classic analyses, enabling new analysis paradigms such as anomaly detection and unfolding, and shifting from predefined features to learned latent representations—and identifies three dangers: statistical shortcuts, ML-defined 'theory' models, and an inflation of low-novelty AI-accelerated publications. It also argues the reverse direction, that particle physics poses challenges to AI, including calibrated class probabilities and bias control, which can drive ML research forward. If the paper is right, the LHC-era adoption of AI will happen without eroding the field's discovery standards.

What carries the argument

The argument is carried by a three-way classification of how ML enters particle physics: ML-enhanced analyses (accelerating and improving existing steps such as triggering, calibration, and event generation), ML-enabled analyses (new paradigms such as weakly supervised anomaly detection and ML unfolding), and representation-based analyses (the transition from predefined feature spaces to learned latent representations). The load-bearing constraint is that learned representations must ultimately be connected to the theory-driven representations defined by the simulation chain, which the paper models as a sequence of factorized conditional probabilities anchored in Lagrangian parameters and symmetries. The two pillars—controlled statistical inference and generalizing theory interpretation—are the filters through which all three directions must pass.

What would settle it

A concrete test is to train a highly expressive ML model on high-energy collision data to predict a measured distribution and then check whether the learned representation can be reproduced by an effective Lagrangian with arbitrarily many higher-dimensional operators, after accounting for detector effects; if an irreducible residual structure survives cross-checks on independent datasets, the claim that ML will not trigger a paradigm shift away from quantum field theory is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim, stated in its own words, is that 'ML developments do not replace the role of theoretical structures, but extend the set of tools through which they can be connected to data.' The authors hold that particle physics is ultimately described by a quantum field theory encoded in a common Lagrangian, so the goal of discovering new physics can be phrased as extracting that Lagrangian from data, and ML's proper role is to provide near-optimal representations and statistically controlled inference strategies for that extraction. They do not expect ML methods by themselves to trigger a paradigm shift away from quantum field theory; a discovery still has to meet the field's statistical requirements and be accompanied by a generalizing theory prediction. The 'vice versa' part of the thesis is that the unusually stringent demands of particle physics—learning class probabilities, controlling biases in learned representations, quantifying uncertainty in high-dimensional spaces—pose questions that go beyond the usual ML scope and can usefully challenge AI research.

Load-bearing premise

The essay assumes that quantum field theory, with its common Lagrangian, remains the correct and sufficient framework for all known fundamental physics, so machine learning can only extend the tools connecting data to this theory and cannot by itself trigger a paradigm shift.

Editorial extensions

If this is right

  • Weakly supervised anomaly searches will be embedded in the same statistical framework as classic bump hunts, so an ML-found signal counts as a discovery only with an understood background model and a look-elsewhere correction.
  • Learned latent representations will be benchmarked against theory-defined objects such as jets, parton densities, and particle flow, making equivariant architectures that exploit known symmetries the default for collider analyses.
  • Agentic systems with access to domain-specific tools—event generators, likelihoods, detector simulations—will become part of the standard workflow, while general-purpose LLM agents alone will not be accepted as a source of physics knowledge.
  • Particle physics will continue to prioritize its statistical foundations, so global versus local significances and calibrated uncertainties will be applied to ML-based analyses, restraining the flood of low-novelty AI-accelerated publications.
  • Physics-driven challenges—calibrated class probabilities, finite-statistics limits of generative models, bias control in network training—will push machine learning research beyond standard benchmarks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the field adopts this position as a norm, ML models whose internal representations cannot be mapped onto quantum-field-theory-based objects will be systematically deprioritized for discovery claims, which would slow the adoption of fully black-box approaches in favor of physics-grounded ones.
  • The quantum-field-theory-anchored assumption implies a testable prediction: a learned latent space trained on hadron-collision data should be reducible to quark/gluon and symmetry-based degrees of freedom; an irreducible topological sector would contradict the paper's assumption.
  • The 'vice versa' direction suggests a concrete benchmark: evaluating uncertainty estimates on class probabilities using collider-style data would sharpen both physics analyses and ML calibration methods, a cross-fertilization the essay points to but does not develop.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. This essay argues that machine learning should be integrated into fundamental physics as a tool for connecting data to theoretical structures, while preserving what it identifies as the two pillars of the field: controlled statistical inference and generalizing theory interpretation. It surveys ML-enhanced and ML-enabled analyses, representation learning, agentic research workflows, and physics-inspired challenges to AI, and closes with recommendations about university training and resource use. The paper is explicitly an essay rather than a technical contribution, and it repeatedly acknowledges open questions and limitations.

Significance. If the conceptual framework holds, the paper provides a useful articulation of a widely held but rarely stated position: ML is a methodological extension that should not replace quantum field theory as the language of fundamental physics. The authors are appropriately cautious, explicitly flagging assumptions (e.g., the QFT framework in Section 2 and the open status of the 'Physics for ML' direction in Section 3). The essay contains no quantitative derivations, code, or machine-checked claims, so its value lies in clarity and framing. The main obstacle to the central claim is the unresolved boundary in Section 2.4 between agent-generated hypotheses and the assertion that ML does not produce theory; this needs to be addressed before the essay's central message is fully coherent.

major comments (2)
  1. [Section 2.4] The description of tool-augmented agents says they can navigate the analysis chain 'from generating hypotheses and producing simulated data to comparing predictions with measurements and updating model parameters.' Section 1, by contrast, says ML 'does not replace the role of theoretical structures, but extend[s] the set of tools through which they can be connected to data.' Hypothesis generation is a theory-construction activity: if an agent proposes a new Lagrangian, a new symmetry, or a new particle content, then ML is not merely connecting data to an existing theoretical structure. The essay gives no criterion for distinguishing legitimate agent-generated hypotheses (which would be evaluated under the two pillars of Section 4) from the 'ML-defined theory models' it explicitly warns against. Please either restrict 'generating hypotheses' to hypothesis tests within a fixed Lagrangian framework or explain how broader agent-generated proposals remain consistent with the claim that ML extends rather than replaces theoretical structures.
  2. [Section 3 / title] The title promises a two-way relation ('and Vice Versa'), but Section 3 states only that particle physics questions 'can, but do not have to inspire research in the direction Physics for ML,' and Section 5 calls this direction 'an interesting question.' The body of the essay therefore does not substantiate the second half of the title. The authors should either moderate the title or provide concrete examples where physics-driven requirements have already produced genuine ML advances, rather than merely stating that such advances are possible.
minor comments (4)
  1. [Section 2.4] The sentence 'agentic systems orchestrate sequences of tasks' should use the plural verb 'orchestrate' rather than 'orchestrates'.
  2. [Section 2] The term 'digital twins' is introduced without a definition; a brief explanation would help readers outside the simulation-based-inference community.
  3. [References] Several of the works cited as evidence for the success of ML methods are authored by the same authors (e.g., refs. 17, 23, 27, 34, 35, 40); the argument would be strengthened by citing more independent examples, especially in Sections 2.1 and 2.4.
  4. [Section 4] The subsection 'University environment' is thematically useful but somewhat disconnected from the preceding technical discussion; a brief transition would improve the flow.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the essay makes a methodological argument without deriving quantitative predictions, and its self-citations are illustrative rather than load-bearing.

full rationale

This paper is a perspective essay rather than a derivation, so the circularity patterns that apply to quantitative papers are largely inapplicable. The central claim is that ML methods should be integrated into fundamental physics while preserving controlled statistical inference and generalizing theory interpretation. The paper repeatedly cites its own prior work (e.g., refs. 17, 23, 27, 34, 35, 40), but these citations are used as examples of existing ML applications in particle physics, not as premises from which the essay's conclusion is derived. No equation is fitted and then renamed as a prediction, no uniqueness theorem is imported from the authors' prior work to force a choice, and no ansatz is smuggled in via citation. The essay explicitly leaves open whether learned latent representations will connect to QFT, stating that this is 'an open and exciting problem,' which further indicates that the argument does not reduce to a self-referential definition. The skeptical concern about Section 2.4 is a substantive tension about whether tool-augmented agents 'generating hypotheses' could cross the paper's own boundary between connecting data to theory and constructing theory, but this is a conceptual ambiguity rather than a circularity: the essay's conclusion is not equivalent to its inputs by construction, and the tension does not make the argument self-validating. Accordingly, the appropriate finding is a low score reflecting only the normal presence of self-citations in a field-review essay, with no load-bearing circular step.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper is an essay; it introduces no fitted parameters and no new entities. Its argument rests on background assumptions about the permanence of the QFT framework, the preservation of statistical standards, and the societal role of physics training, all stated explicitly in Sections 2 and 4.

assumptions (4)
  • domain assumption Quantum field theory, encoded in a Lagrangian or action, is the correct and sufficient framework for describing all known fundamental physics, and ML by itself will not trigger a paradigm shift.
    Invoked in Section 2, where new-physics discovery is framed as extracting the Lagrangian; the authors explicitly say they do not expect ML to change this framework.
  • domain assumption The statistical discovery standards of particle physics, including controlled inference, hypothesis tests, and look-elsewhere control, should be preserved unchanged when ML is integrated.
    Section 4 states 'AI should not weaken the two pillars of fundamental physics: controlled statistical inference and generalizing theory interpretation'; this normative premise is the basis of the paper's recommendations.
  • domain assumption The societal value of university fundamental physics comes substantially from training students and future leaders, not only from research output.
    Section 4 (University environment) argues that educational mission must be part of the AI strategy; this assumption motivates the call to embed AI training.
  • domain assumption The AI transformation of society and research is faster than historical precedents and leaves little adaptation time.
    Section 5 states this as a fact without quantification; it drives the paper's sense of urgency.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Machine Learning is Good for Physics - and Vice Versa." pith.science (2026). https://pith.science/paper/6UINPQ6X

@misc{pith2026260805812,
  author       = {Pith},
  title        = {Pith review of: Machine Learning is Good for Physics - and Vice Versa},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6UINPQ6X}},
  note         = {Machine review of arXiv:2608.05812}
}
read the original abstract

Scientific AI is rapidly transforming fundamental physics research and challenging defining aspects of the fundamental physics methodology. We discuss opportunities and dangers of this transformation and find exciting benefits from a close interaction between AI and fundamental physics, provided that we remain aware of the scientific methodologies of the respective fields. For fundamental physics, we discuss two such aspects: statistical validation and a generalizing theory description, both with the goal of discovering new physics in vast datasets.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 5 canonical work pages

  1. [1]

    Trotta,The indiscriminate adoption of ai threatens the foundations of academia, Nature Astronomy9(12), 1748–1749 (2025), doi:10.1038/s41550-025-02738-w

    R. Trotta,The indiscriminate adoption of ai threatens the foundations of academia, Nature Astronomy9(12), 1748–1749 (2025), doi:10.1038/s41550-025-02738-w

    doi:10.1038/s41550-025-02738-w 2025

  2. [2]

    D. W. Hogg and S. Villar,Is machine learning good or bad for the natural sciences?, arXiv e- prints arXiv:2405.18095 (2024), doi:10.48550/arXiv. 2405.18095,2405.18095

    doi:10.48550/arxiv.2405.18095 2024

  3. [3]

    Are We Ready for AI-Driven Discovery? AI Verification Before the Next Fundamental Physics Breakthrough

    G. Grosso, V. Mikuni and L. Heinrich,Are We Ready for AI-Driven Discovery? AI Verification Before the Next Fundamental Physics Breakthrough(2026),2607.10039

    work page Pith review arXiv 2026

  4. [4]

    Manyika (editor),Ai & science: What is the future of discovery?, DaedalusWinter/Spring(2026)

    James M. Manyika (editor),Ai & science: What is the future of discovery?, DaedalusWinter/Spring(2026)

    work page 2026

  5. [5]

    Feyerabend,Against Method, New Left Books, London and New York (1975)

    P. Feyerabend,Against Method, New Left Books, London and New York (1975)

    work page 1975

  6. [6]

    Srednicki,Quantum field theory, Cambridge Univer- sity Press, ISBN 978-0-521-86449-7, 978-0-511-26720-8, doi:10.1017/CBO9780511813917 (2007)

    M. Srednicki,Quantum field theory, Cambridge Univer- sity Press, ISBN 978-0-521-86449-7, 978-0-511-26720-8, doi:10.1017/CBO9780511813917 (2007)

    doi:10.1017/cbo9780511813917 2007

  7. [7]

    M. D. Schwartz,Quantum Field Theory and the Standard Model, Cambridge University Press, ISBN 978-1-107-03473-0, 978-1-107-03473-0, doi:10.1017/ 9781139540940 (2014)

    work page 2014

  8. [8]

    Cranmer, J

    K. Cranmer, J. Brehmer and G. Louppe,The fron- tier of simulation-based inference, Proc. Nat. Acad. Sci. 117(48), 30055 (2020), doi:10.1073/pnas.1912789117, 1911.01429

    arXiv 2020

Show all 44 references
  1. [9]

    ATLAS Collaboration,A measurement of the high-mass τ τproduction cross-section at √s= 13TeV with the AT- LAS detector and constraints on new particles and cou- plings, JHEP10, 054 (2025), doi:10.1007/JHEP10(2025) 054,2503.19836

    2025

  2. [10]

    de Oliveira, M

    L. de Oliveira, M. Kagan, L. Mackey, B. Nachman and A. Schwartzman,Jet-images — deep learning edition, JHEP07, 069 (2016), doi:10.1007/JHEP07(2016)069, 1511.05190

    2016 arXiv

  3. [11]

    P. T. Komiske, E. M. Metodiev and J. Thaler,Energy Flow Networks: Deep Sets for Particle Jets, JHEP01, 121 (2019), doi:10.1007/JHEP01(2019)121,1810.05165

    2019 arXiv

  4. [12]

    Qu and L

    H. Qu and L. Gouskos,ParticleNet: Jet Tagging via Particle Clouds, Phys. Rev. D101(5), 056019 (2020), doi:10.1103/PhysRevD.101.056019,1902.08570

    2020 arXiv

  5. [13]

    Butteret al.,The Machine Learning landscape of top taggers, SciPost Phys.7, 014 (2019), doi:10.21468/ SciPostPhys.7.1.014,1902.09914

    A. Butteret al.,The Machine Learning landscape of top taggers, SciPost Phys.7, 014 (2019), doi:10.21468/ SciPostPhys.7.1.014,1902.09914

    2019 arXiv

  6. [14]

    J. M. Campbellet al.,Event generators for high-energy physics experiments, SciPost Phys.16(5), 130 (2024), doi:10.21468/SciPostPhys.16.5.130,2203.11110

    2024 arXiv

  7. [15]

    Badgeret al.,Machine learning and LHC event gen- eration, SciPost Phys.14(4), 079 (2023), doi:10.21468/ SciPostPhys.14.4.079,2203.07460

    S. Badgeret al.,Machine learning and LHC event gen- eration, SciPost Phys.14(4), 079 (2023), doi:10.21468/ SciPostPhys.14.4.079,2203.07460

    2023 arXiv

  8. [16]

    Janßen, R

    T. Janßen, R. Poncelet and S. Schumann,Sampling NNLO QCD phase space with normalizing flows, JHEP 09, 194 (2025), doi:10.1007/JHEP09(2025)194,2505. 13608

    2025 doi

  9. [17]

    De Crescenzo, J

    G. De Crescenzo, J. M. Villadamigo, N. Elmer, T. Heimel, T. Plehn, R. Winterhalder and M. Zaro,Mad- NIS at NLO(2026),2603.22407

    2026 arXiv

  10. [18]

    Nachman and D

    B. Nachman and D. Shih,Anomaly Detection with Den- sity Estimation, Phys. Rev. D101, 075042 (2020), doi: 10.1103/PhysRevD.101.075042,2001.04990

    2020 arXiv

  11. [19]

    Hallin, J

    A. Hallin, J. Isaacson, G. Kasieczka, C. Krause, B. Nach- man, T. Quadfasel, M. Schlaffer, D. Shih and M. Som- merhalder,Classifying anomalies through outer density estimation, Phys. Rev. D106(5), 055006 (2022), doi: 10.1103/PhysRevD.106.055006,2109.00546

    2022 arXiv

  12. [20]

    M. Hein, B. Nachman and D. Shih,Look everywhere effects in anomaly detection(2025),2512.13787

    2025 arXiv

  13. [21]

    Heimel, G

    T. Heimel, G. Kasieczka, T. Plehn and J. M. Thompson, QCD or What?, SciPost Phys.6(3), 030 (2019), doi: 10.21468/SciPostPhys.6.3.030,1808.08979

    2019 arXiv

  14. [22]

    Farina, Y

    M. Farina, Y. Nakai and D. Shih,Searching for New Physics with Deep Autoencoders, Phys. Rev. D101(7), 075021 (2020), doi:10.1103/PhysRevD.101.075021,1808. 08992

    2020 doi

  15. [23]

    B. M. Dillon, L. Favaro, T. Plehn, P. Sorrenson and M. Kr¨ amer,A normalized autoencoder for LHC trig- gers, SciPost Phys. Core6, 074 (2023), doi:10.21468/ SciPostPhysCore.6.4.074,2206.14225

    2023

  16. [24]

    Andreassen, P

    A. Andreassen, P. T. Komiske, E. M. Metodiev, B. Nach- man and J. Thaler,OmniFold: A Method to Simul- taneously Unfold All Observables, Phys. Rev. Lett. 124(18), 182001 (2020), doi:10.1103/PhysRevLett.124. 182001,1911.09107

    2020 arXiv

  17. [25]

    Bellagente, A

    M. Bellagente, A. Butter, G. Kasieczka, T. Plehn, A. Rousselot, R. Winterhalder, L. Ardizzone and U. K¨ othe,Invertible Networks or Partons to Detector and Back Again, SciPost Phys.9, 074 (2020), doi: 10.21468/SciPostPhys.9.5.074,2006.06685

    2020 arXiv

  18. [26]

    Bengio, A

    Y. Bengio, A. C. Courville and P. Vincent,Unsupervised feature learning and deep learning: A review and new perspectives, CoRRabs/1206.5538(2012),1206.5538

    2012 arXiv

  19. [27]

    Brehmer, V

    J. Brehmer, V. Bres´ o, P. de Haan, T. Plehn, H. Qu, J. Spinner and J. Thaler,A Lorentz-equivariant trans- former for all of the LHC, SciPost Phys.19(4), 108 (2025), doi:10.21468/SciPostPhys.19.4.108,2411.00446

    2025

  20. [28]

    Favaro, G

    L. Favaro, G. Gerhartz, F. A. Hamprecht, P. Lippmann, S. Pitz, T. Plehn, H. Qu and J. Spinner,Lorentz- Equivariance without Limitations(2025),2508.14898

    2025

  21. [29]

    M. M. Bronstein, J. Bruna, T. Cohen and P. Velick- ovic,Geometric deep learning: Grids, groups, graphs, geodesics, and gauges, CoRRabs/2104.13478(2021), 2104.13478

    2021 arXiv

  22. [30]

    Golling, L

    T. Golling, L. Heinrich, M. Kagan, S. Klein, M. Leigh, M. Osadchy and J. A. Raine,Masked particle model- ing on sets: towards self-supervised high energy physics foundation models, Mach. Learn. Sci. Tech.5(3), 035074 (2024), doi:10.1088/2632-2153/ad64a8,2401.13537

    2024 arXiv

  23. [31]

    J. Birk, A. Hallin and G. Kasieczka,OmniJet-α: the first cross-task foundation model for particle physics, Mach. Learn. Sci. Tech.5(3), 035031 (2024), doi:10.1088/ 2632-2153/ad66ad,2403.05618. 10

    2024 arXiv

  24. [32]

    Mikuni and B

    V. Mikuni and B. Nachman,Solving key challenges in collider physics with foundation models, Phys. Rev. D111(5), L051504 (2025), doi:10.1103/PhysRevD.111. L051504,2404.16091

    2025

  25. [33]

    Bommasani, D

    R. Bommasani, D. A. Hudson, E. Adeli, R. B. Altman, S. Arora, S. von Arx, M. S. Bernstein, J. Bohg, A. Bosse- lut, E. Brunskill, E. Brynjolfsson, S. Buchet al.,On the opportunities and risks of foundation models, CoRR abs/2108.07258(2021),2108.07258

    2021 arXiv

  26. [34]

    Diefenbacher, A

    S. Diefenbacher, A. Hallin, G. Kasieczka, M. Kr¨ amer, A. Lauscher and T. Lukas,Agents of Discovery(2025), 2509.08535

    2025

  27. [35]

    Plehn, D

    T. Plehn, D. Schiller and N. Schmal,MadAgents(2026), 2601.21015

    2026 arXiv

  28. [36]

    E. M. Metodiev, B. Nachman and J. Thaler,Clas- sification without labels: Learning from mixed samples in high energy physics, JHEP10, 174 (2017), doi: 10.1007/JHEP10(2017)174,1708.02949

    2017 arXiv

  29. [37]

    Butter, S

    A. Butter, S. Diefenbacher, G. Kasieczka, B. Nachman and T. Plehn,GANplifying event samples, SciPost Phys. 10(6), 139 (2021), doi:10.21468/SciPostPhys.10.6.139, 2008.06545

    2021 arXiv

  30. [38]

    R. D. Ballet al.,The path to N 3LO parton distributions, Eur. Phys. J. C84(7), 659 (2024), doi:10.1140/epjc/ s10052-024-12891-7,2402.18635

    2024 arXiv

  31. [39]

    Arvanitidis, L

    G. Arvanitidis, L. K. Hansen and S. Hauberg,Latent space oddity: on the curvature of deep generative models (2021),1710.11379

    2021 arXiv

  32. [40]

    R. M. Kuntz, T. Plehn, B. M. Sch¨ afer, B. Schosser and S. Vent,The Latent Information Geometry of Jet Clas- sification(2026),2603.02310

    2026 arXiv

  33. [41]

    A. Bal, M. Klute, B. Maier and M. Spannowsky,From Information Geometry to Jet Substructure: A Triality of Cumulant Tensors, Energy Correlators, and Hyper- graphs(2026),2605.03063

    2026 arXiv

  34. [42]

    D. A. Roberts, S. Yaida and B. Hanin,The Principles of Deep Learning Theory, Cambridge University Press, ISBN 978-1-009-02340-5, doi:10.1017/9781009023405 (2022),2106.10165

    2022 arXiv

  35. [43]

    Ringel, N

    Z. Ringel, N. Rubin, E. Mor, M. Helias and I. Seroussi, Applications of statistical field theory in deep learning (2025),2502.18553

    2025 arXiv

  36. [44]

    Luccioni, Y

    S. Luccioni, Y. Jernite and E. Strubell,Power hungry processing: Watts driving the cost of ai deployment?, In The 2024 ACM Conference on Fairness Accountability and Transparency, F AccT ’24, p. 85–99. ACM, doi:10. 1145/3630106.3658542 (2024)

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.