Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Beyond Equilibrium: Non-Equilibrium Foundations Should Underpin Generative Processes in Complex Dynamical Systems

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read Non-equilibrium statistical mechanics, not equilibrium assumptions, is the necessary foundation for generative models of time-evolving complex systems, this position paper argues.

desk verdict A well-written survey and position piece whose 'necessary' claim rests on a single confounded 2D experiment; worth reviewing but only after the evidence is brought in line with the rhetoric. read the letter →

arxiv 2505.18621 v1 pith:OHAA4H7X submitted 2025-05-24 cs.CE

classification cs.CE
keywords non-equilibriumstatisticalmechanicsgenerativemodelsdiffusionenergy-basedcomplexdynamicalsystemstime-varyingpotentialsLangevindynamicsscore-basedmodeling
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

This position paper claims that generative models grounded in non-equilibrium statistical mechanics—not equilibrium-based models—are the necessary foundation for simulating, understanding, and controlling time-evolving complex systems. The authors argue that equilibrium-inspired methods such as energy-based models and Boltzmann sampling assume a static distribution, so they cannot represent transient, irreversible, or far-from-equilibrium dynamics. To support the position, they simulate particles in a time-varying Printz potential with overdamped Langevin dynamics, and compare an equilibrium method that fits a static energy field and samples via Boltzmann with a non-equilibrium conditional diffusion model that learns the time-dependent score and samples via denoising Langevin dynamics. The non-equilibrium method consistently achieves lower Jensen-Shannon divergence to the true evolving particle distribution. If correct, the claim reframes diffusion-style generative models as the default scientific tool for dynamical systems rather than a convenient alternative.

What carries the argument

The load-bearing device is the contrast between two ways of modeling the same time-varying system. The test system is a two-dimensional Printz potential $V(t,x,y) = \cos(s\arctan(y,x) - \frac{\pi}{2}t) + 10(\sqrt{x^2+y^2}-\frac12)^2$ with $s=5$, evolving under overdamped Langevin dynamics $dX_t = -\nabla V(X_t,t)\,dt + \sqrt{2\beta^{-1}}\,dW_t$. The equilibrium strategy distills the instantaneous energy field into a neural field $V_\theta(t,x,y)$ and samples from the Boltzmann distribution $p \propto e^{-V}$; the non-equilibrium strategy learns the conditional score function $\nabla_x \log p(x|t)$ and samples by denoising Langevin dynamics. The paper's argument turns on the score/gradient-field representation: the non-equilibrium model tracks the energy landscape's evolution directly, whereas the equilibrium model must re-estimate a static field at each time.

What would settle it

Run the same rotating Printz potential comparison with the model family held fixed: train the same neural architecture once with a time-varying Boltzmann objective and once with score matching, and compare the Jensen-Shannon divergence of generated particle distributions. If the gap vanishes or reverses, the claim that non-equilibrium foundations are necessary for such systems would not be established by this experiment.

Watch

Extended reading notes

Core claim

The paper's central claim is that embracing non-equilibrium physics is not merely beneficial but necessary for generative AI to serve as a scientific modeling tool for complex dynamical systems. On the paper's terms, equilibrium-inspired generative models inherit the Boltzmann distribution's stationarity assumption, and therefore fail to represent systems that are transient, irreversible, or far from equilibrium; non-equilibrium-inspired generative models, by contrast, treat generation as a Markov-chain or stochastic-process transformation that tracks time-dependent energy gradients. The experimental demonstration on a two-dimensional Printz potential with a rotating energy landscape shows that a conditional diffusion model using denoising Langevin dynamics gives lower Jensen-Shannon divergence for time-dependent particle distributions than an equilibrium approach that fits a time-varying energy field and performs Boltzmann sampling. The paper also argues that non-equilibrium frameworks naturally encode entropy production, irreversible flows, and fluctuation theorems, opening routes to rare-event simulation, mechanism inference, and multi-scale modeling.

Load-bearing premise

The load-bearing premise is that the observed performance gap comes from the equilibrium versus non-equilibrium distinction itself, not from the different model families—a neural conditional diffusion model against a histogram-based static energy field with Boltzmann sampling—and that the Gaussian noise in the Langevin simulator is irrelevant to the comparison.

Editorial extensions

If this is right

  • If the position is correct, diffusion-style non-equilibrium generative models become the default foundation for modeling real-world dynamical systems, not a niche alternative to equilibrium energy-based models.
  • Equilibrium energy-based models will remain limited for systems that lack a well-defined or stationary energy landscape, including transient and far-from-equilibrium regimes.
  • Conditional score-based diffusion can generate time-dependent particle distributions that track evolving gradients, as demonstrated on the rotating Printz potential.
  • Non-equilibrium generative models offer a route to rare-event simulation, mechanism inference, and multi-scale modeling by encoding entropy production, irreversible flows, and fluctuation theorems.
  • The outlook implies generative AI can move beyond static distribution fitting to modeling the full unfolding of irreversible, history-dependent processes.

Reading between the lines

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

  • Holding the model family fixed would isolate the equilibrium/non-equilibrium distinction: train the same neural architecture once with a time-varying Boltzmann objective and once with score matching, then compare JSD on the same rotating potential; the paper's necessity claim would be sharpened by that experiment.
  • Because score-based diffusion is itself a non-equilibrium process, the practical novelty of the position may lie less in inventing new generators and more in explicitly conditioning existing diffusion models on time, evolving energy fields, and physical priors such as entropy production.
  • A concrete extension suggested by the outlook is to regularize conditional diffusion training with entropy-production or fluctuation-theorem terms and measure whether tracking of transient distributions improves on the Printz system.
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

4 major / 5 minor

Summary. This position paper argues that generative models grounded in equilibrium statistical mechanics are fundamentally limited for modeling complex dynamical systems, and that non-equilibrium-inspired generative models are not merely helpful but necessary for scientific generative AI. The paper surveys equilibrium-based models (EBMs, RBMs, Hopfield networks), non-equilibrium-inspired models (diffusion, Schrödinger bridges, Poisson flow), and other generative families (flows, VAEs, autoregressive models, GANs), connecting them to statistical physics concepts. It reports an experiment on a two-dimensional time-varying Printz potential system, comparing an equilibrium baseline (histogram-derived static energy field distilled into a neural field, followed by Boltzmann sampling) with a non-equilibrium baseline (conditional score-based diffusion model). The authors claim the non-equilibrium method achieves consistently lower Jensen-Shannon divergence for generated particle distributions. The paper closes with a broad set of proposed future directions for non-equilibrium physics in generative AI.

Significance. If the central necessity claim were established, the paper would provide a strong argument for reorienting generative modeling for scientific simulation toward non-equilibrium frameworks. The paper offers a useful synthesis of existing physics-inspired generative models and a concrete, minimal benchmark in a time-varying potential. However, the empirical evidence is currently too confounded to support the strong claim of necessity: the two compared methods differ in model family, expressiveness, training objective, and noise-structure matching, and the experiment lacks error bars, ablations, and statistical testing. The paper is most valuable as a position piece and research roadmap; as a demonstration of necessity, it is not yet convincing.

major comments (4)
  1. [Section 3 / Appendix D] The central comparison conflates the equilibrium/non-equilibrium distinction with model family and training objective. The equilibrium baseline is a histogram-derived static energy field distilled into a neural field Vθ followed by Boltzmann sampling, while the non-equilibrium baseline is a conditional score-based diffusion model with a neural score function. These methods differ in expressiveness, parametrization, and loss, so the lower JSD reported in Figure 1b could be due to any of these factors rather than the physical principle under test. To support the necessity claim, please add a controlled comparison: an equilibrium-inspired method with the same neural architecture and training procedure (e.g., learning a time-dependent energy field with score matching or denoising, then sampling via Langevin dynamics), and a non-equilibrium method with a histogram-based or otherwise restricted score estimator.
  2. [Appendix D, Eq. (31)] The data-generating process is an overdamped Langevin SDE with additive Gaussian white noise, and the non-equilibrium generative model is a score-based diffusion model whose forward process injects Gaussian noise. The non-equilibrium method is therefore matched to the simulator's noise structure by construction, making the experiment in part a self-consistency check of Gaussian diffusion on a Gaussian-noise process. To rule out this alternative explanation, test the comparison under non-Gaussian perturbations (e.g., Lévy noise or multiplicative noise) or with a different generative forward process, and show that the non-equilibrium method still outperforms the equilibrium baseline.
  3. [Figure 1b / Appendix D] The paper reports no quantitative values, error bars, number of seeds, or statistical tests for the JSD comparison, and Appendix D omits key hyperparameters such as histogram bin width, diffusion steps, network architecture, and training details. The claim that the non-equilibrium method is 'consistently' lower cannot be verified without this information. Please report mean and standard deviation over multiple independent trajectory sets and training seeds, and provide the full experimental configuration for reproducibility.
  4. [Abstract / Section 1] The paper's central assertion is that non-equilibrium frameworks are 'necessary' for generative AI in science. Even if the proposed experiments were controlled, a single two-dimensional synthetic system with one equilibrium baseline and one non-equilibrium baseline would at most establish that the non-equilibrium approach is beneficial in that setting. Please either soften the necessity claim to a position or research hypothesis, or provide substantially broader evidence, including equilibrium methods with matched capacity and multiple systems, to justify the stronger statement.
minor comments (5)
  1. [Appendix B] There is a typo in Appendix B: 'whicn' should be 'which'.
  2. [Section 4.1] The word 'proterns' in the multi-scale systems paragraph should be 'proteins'.
  3. [Appendix A] The heading 'Diffussion Schödinger Bridge Model' contains a typo: 'Diffussion' should be 'Diffusion', and 'Schödinger' should be 'Schrödinger'.
  4. [Section 2.2 / Appendix B] The text says the Fokker-Planck equation derivation is in Appendix A, but the derivation actually appears in Appendix B; the cross-reference should be corrected.
  5. [Appendix A] Table 2 abbreviates the 'Diffusion Schrödinger Bridge Model' as 'SBM', but this abbreviation is not defined and is used only in the table; please define it or use a consistent name.

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity: the confounded experiment and self-citations do not reduce the paper's claims to their inputs by construction.

full rationale

The paper's central claim is a position statement, not a derivation from first principles, so most of the text is interpretive rather than circular. The only empirical load-bearing component is Section 3 and Appendix D, which compares a conditional score-based diffusion model against a histogram-distilled Boltzmann energy field on a time-varying Printz potential. The physical simulator in Eq. (31) and the conditional diffusion model both involve Gaussian noise, and the comparison does not isolate the equilibrium/non-equilibrium distinction: model family, training objective, expressiveness, and noise-structure matching all differ between the two methods. This is a serious validity threat to the 'necessary' claim, but it is not circularity. The non-equilibrium model's success is not equated to its input by definition; the score network must be trained, the reverse SDE must be sampled, and the JSD gap is an empirical outcome rather than a fitted quantity renamed as a prediction. Reference [79], which supplies the test system, is a self-citation, but it is not load-bearing for the position claim, and no uniqueness theorem or ansatz is imported from it. The mathematical links in Section 2 (e.g., ELBO paralleling variational free energy, diffusion as a non-equilibrium Markov chain) are analogies and identifications, not reductions of the conclusion to an assumption. Therefore, no significant circularity is present; the appropriate criticism belongs to experimental validity and generalizability, not circularity.

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

The ledger is small because the paper is a position piece, not a parameterized theory. The main uncharged assumptions are the representativeness of the toy system and the fairness of the equilibrium/non-equilibrium comparison.

free parameters (4)
  • s (number of potential wells) = s = 5
    Hand-chosen in the simulation (Appendix D) to create a five-well rotating landscape; not derived or fitted, but the demo's difficulty and results depend on it.
  • beta (inverse noise intensity) = beta = 10
    Hand-chosen noise level in the overdamped Langevin dynamics (Appendix D); controls how much the non-equilibrium diffusion model matches the data noise.
  • histogram bin width for equilibrium energy field = not reported
    The equilibrium baseline builds a static energy field from sample frequencies; the binning bandwidth is a free parameter that materially affects its quality and is not specified.
  • conditional diffusion model hyperparameters = not reported
    Network architecture, noise schedule, and training budget for the conditional score model are not provided, so the demo cannot be reproduced precisely.
assumptions (4)
  • domain assumption The Printz potential system with overdamped Langevin dynamics is representative of real-world complex dynamical systems.
    Invoked in Section 3 and Appendix D to justify that conclusions from this 2D toy transfer to molecular, climate, and network systems.
  • domain assumption Gaussian white noise in the data-generating SDE matches the Gaussian noise assumption of score-based diffusion models.
    The experiment's non-equilibrium method is matched to the simulator's noise model; this matching is assumed to be incidental rather than the cause of the performance gap.
  • standard math Standard Ito calculus and Fokker-Planck results used in Appendices A-C.
    The Fokker-Planck equation and Ito's lemma are standard background results used without proof.
  • domain assumption Jensen-Shannon divergence adequately measures how well a generative model captures temporal evolution.
    Section 3 uses JSD as the sole error metric; other metrics (e.g., autocorrelation, entropy production, transition probabilities) might give different rankings.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Beyond Equilibrium: Non-Equilibrium Foundations Should Underpin Generative Processes in Complex Dynamical Systems." pith.science (2026). https://pith.science/paper/OHAA4H7X

@misc{pith2026250518621,
  author       = {Pith},
  title        = {Pith review of: Beyond Equilibrium: Non-Equilibrium Foundations Should Underpin Generative Processes in Complex Dynamical Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OHAA4H7X}},
  note         = {Machine review of arXiv:2505.18621}
}
read the original abstract

This position paper argues that next-generation non-equilibrium-inspired generative models will provide the essential foundation for better modeling real-world complex dynamical systems. While many classical generative algorithms draw inspiration from equilibrium physics, they are fundamentally limited in representing systems with transient, irreversible, or far-from-equilibrium behavior. We show that non-equilibrium frameworks naturally capture non-equilibrium processes and evolving distributions. Through empirical experiments on a dynamic Printz potential system, we demonstrate that non-equilibrium generative models better track temporal evolution and adapt to non-stationary landscapes. We further highlight future directions such as integrating non-equilibrium principles with generative AI to simulate rare events, inferring underlying mechanisms, and representing multi-scale dynamics across scientific domains. Our position is that embracing non-equilibrium physics is not merely beneficial--but necessary--for generative AI to serve as a scientific modeling tool, offering new capabilities for simulating, understanding, and controlling complex systems.

Figures

Figures reproduced from arXiv: 2505.18621 by the authors.

Figure 1
Figure 1. Equilibrium vs. non-equilibrium generation on a time-varying potential system. (a) [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Outlook or future directions of both non-equilibrium physics and generative AI. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Physical systems invoke non-equilibrium thermodynamics and the associated problems. [PITH_FULL_IMAGE:figures/full_fig_p026_3.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Quantum Reversibility Meets Classical Reverse Diffusion

    quant-ph 2025-10 conditional novelty 4.0 of 10

    The semiclassical limit of the Petz-reversed Lindblad equation reproduces the Bayes-rule reverse-time diffusion equation, with the reference state's Wigner function playing the role of the classical score distribution.

Reference graph

Works this paper leans on

180 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [1]

    Information theory and statistical mechanics

    Edwin T Jaynes. Information theory and statistical mechanics. Physical review, 106(4):620, 1957

  2. [2]

    Weitere studien über das wärmegleichgewicht unter gasmolekülen, vol- ume 66

    Ludwig Boltzmann. Weitere studien über das wärmegleichgewicht unter gasmolekülen, vol- ume 66. Aus der kk Hot-und Staatsdruckerei, 1872

  3. [3]

    Elementary principles in statistical mechanics: developed with especial reference to the rational foundations of thermodynamics

    Josiah Willard Gibbs. Elementary principles in statistical mechanics: developed with especial reference to the rational foundations of thermodynamics. C. Scribner’s sons, 1902

  4. [4]

    Neural networks and physical systems with emergent collective computational abilities

    John J Hopfield. Neural networks and physical systems with emergent collective computational abilities. Proceedings of the national academy of sciences, 79(8):2554–2558, 1982

  5. [5]

    Carnegie-Mellon University, Department of Computer Science Pittsburgh, PA, 1984

    Geoffrey E Hinton, Terrence J Sejnowski, and David H Ackley.Boltzmann machines: Con- straint satisfaction networks that learn. Carnegie-Mellon University, Department of Computer Science Pittsburgh, PA, 1984

  6. [6]

    A learning algorithm for boltzmann machines

    David H Ackley, Geoffrey E Hinton, and Terrence J Sejnowski. A learning algorithm for boltzmann machines. Cognitive science, 9(1):147–169, 1985

  7. [7]

    Learning and relearning in boltzmann machines

    Geoffrey E Hinton, Terrence J Sejnowski, et al. Learning and relearning in boltzmann machines. Parallel distributed processing: Explorations in the microstructure of cognition, 1(282-317):2, 1986

  8. [8]

    Training products of experts by minimizing contrastive divergence.Neural computation, 14(8):1771–1800, 2002

    Geoffrey E Hinton. Training products of experts by minimizing contrastive divergence.Neural computation, 14(8):1771–1800, 2002

Show all 180 references
  1. [9]

    A tutorial on energy-based learning

    Yann LeCun, Sumit Chopra, Raia Hadsell, Marc’Aurelio Ranzato, and Fu Jie Huang. A tutorial on energy-based learning. Predicting structured data, 1(0), 2006

  2. [10]

    Implicit generation and modeling with energy based models

    Yilun Du and Igor Mordatch. Implicit generation and modeling with energy based models. Advances in Neural Information Processing Systems, 32, 2019

  3. [11]

    Your classifier is secretly an energy based model and you should treat it like one

    Will Grathwohl, Kuan-Chieh Wang, Jörn-Henrik Jacobsen, David Duvenaud, Mohammad Norouzi, and Kevin Swersky. Your classifier is secretly an energy based model and you should treat it like one. arXiv preprint arXiv:1912.03263, 2019

  4. [12]

    Statistical mechanics: a set of lectures

    Richard P Feynman. Statistical mechanics: a set of lectures. CRC press, 2018

  5. [13]

    Auto-encoding variational bayes

    Diederik P Kingma and Max Welling. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013

  6. [14]

    Stochastic backpropagation and approximate inference in deep generative models

    Danilo Jimenez Rezende, Shakir Mohamed, and Daan Wierstra. Stochastic backpropagation and approximate inference in deep generative models. In International conference on machine learning, pages 1278–1286, 2014

  7. [15]

    Improved variational inference with inverse autoregressive flow.Advances in neural informa- tion processing systems, 29, 2016

    Durk P Kingma, Tim Salimans, Rafal Jozefowicz, Xi Chen, Ilya Sutskever, and Max Welling. Improved variational inference with inverse autoregressive flow.Advances in neural informa- tion processing systems, 29, 2016

  8. [16]

    Variational diffusion models

    Diederik Kingma, Tim Salimans, Ben Poole, and Jonathan Ho. Variational diffusion models. Advances in neural information processing systems, 34:21696–21707, 2021

  9. [17]

    Poisson variational autoencoder

    Hadi Vafaii, Dekel Galor, and Jacob Yates. Poisson variational autoencoder. Advances in Neural Information Processing Systems, 37:44871–44906, 2024

  10. [18]

    Sur la théorie relativiste de l’électron et l’interprétation de la mécanique quantique

    Erwin Schrödinger. Sur la théorie relativiste de l’électron et l’interprétation de la mécanique quantique. In Annales de l’institut Henri Poincaré, volume 2, pages 269–310, 1932

  11. [19]

    Reciprocal relations in irreversible processes

    Lars Onsager. Reciprocal relations in irreversible processes. ii. Physical review, 38(12):2265, 1931

  12. [20]

    Diffusion schrödinger bridge with applications to score-based generative modeling

    Valentin De Bortoli, James Thornton, Jeremy Heng, and Arnaud Doucet. Diffusion schrödinger bridge with applications to score-based generative modeling. Advances in Neural Information Processing Systems, 34:17695–17709, 2021. 10

  13. [21]

    Deep unsuper- vised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsuper- vised learning using nonequilibrium thermodynamics. In International conference on machine learning, pages 2256–2265. PMLR, 2015

  14. [22]

    Denoising diffusion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models.Advances in neural information processing systems, 33:6840–6851, 2020

  15. [23]

    Score-based generative modeling through stochastic differential equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456, 2020

  16. [24]

    Correlation functions and computer simulations

    Giorgio Parisi. Correlation functions and computer simulations. Nuclear Physics B, 180(3): 378–384, 1981

  17. [25]

    Generative modeling by estimating gradients of the data distribution

    Yang Song and Stefano Ermon. Generative modeling by estimating gradients of the data distribution. Advances in neural information processing systems, 32, 2019

  18. [26]

    Fp-diffusion: Improving score-based diffusion models by enforcing the underlying score fokker-planck equation

    Chieh-Hsin Lai, Yuhta Takida, Naoki Murata, Toshimitsu Uesaka, Yuki Mitsufuji, and Stefano Ermon. Fp-diffusion: Improving score-based diffusion models by enforcing the underlying score fokker-planck equation. In International Conference on Machine Learning , pages 18365–18398....

  19. [27]

    Poisson flow generative models

    Yilun Xu, Ziming Liu, Max Tegmark, and Tommi Jaakkola. Poisson flow generative models. Advances in Neural Information Processing Systems, 35:16782–16795, 2022

  20. [28]

    Diffusion schrödinger bridge matching

    Yuyang Shi, Valentin De Bortoli, Andrew Campbell, and Arnaud Doucet. Diffusion schrödinger bridge matching. Advances in Neural Information Processing Systems, 36, 2024

  21. [29]

    Transport meets variational inference: Controlled monte carlo diffusions

    Francisco Vargas, Shreyas Padhy, Denis Blessing, and Nikolas Nüsken. Transport meets variational inference: Controlled monte carlo diffusions. arXiv preprint arXiv:2307.01050, 2023

  22. [30]

    Escorted free energy simulations: Improving convergence by reducing dissipation

    Suriyanarayanan Vaikuntanathan and Christopher Jarzynski. Escorted free energy simulations: Improving convergence by reducing dissipation. Physical Review Letters, 100(19):190601, 2008

  23. [31]

    Jarzynski

    C. Jarzynski. Nonequilibrium equality for free energy differences. Physical Review Letters, 78(14):2690–2693, 1997

  24. [32]

    Entropy production along a stochastic trajectory and an integral fluctuation theorem

    Udo Seifert. Entropy production along a stochastic trajectory and an integral fluctuation theorem. Physical review letters, 95(4):040602, 2005

  25. [33]

    Density estimation using real nvp

    Laurent Dinh, Jascha Sohl-Dickstein, and Samy Bengio. Density estimation using real nvp. In ICLR, 2017

  26. [34]

    Glow: Generative flow with invertible 1x1 convolutions

    Durk P Kingma and Prafulla Dhariwal. Glow: Generative flow with invertible 1x1 convolutions. In NeurIPS, 2018

  27. [35]

    Ffjord: Free-form continuous dynamics for scalable reversible generative models

    Will Grathwohl, Ricky TQ Chen, Jesse Bettencourt, Ilya Sutskever, and David Duvenaud. Ffjord: Free-form continuous dynamics for scalable reversible generative models. In ICLR, 2019

  28. [36]

    Pixel recurrent neural networks

    Aaron van den Oord, Nal Kalchbrenner, and Koray Kavukcuoglu. Pixel recurrent neural networks. In ICML, 2016

  29. [37]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017

  30. [38]

    Improving language understanding by generative pre-training

    Alec Radford, Karthik Narasimhan, Tim Salimans, Ilya Sutskever, et al. Improving language understanding by generative pre-training. 2018

  31. [39]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In Advances in neural information processing systems, pages 2672–2680, 2014. 11

  32. [40]

    Statistical Physics: Volume 5, volume 5

    Lev Davidovich Landau and Evgenii Mikhailovich Lifshitz. Statistical Physics: Volume 5, volume 5. Elsevier, 2013

  33. [41]

    Restricted boltzmann machines for collaborative filtering

    Ruslan Salakhutdinov, Andriy Mnih, and Geoffrey Hinton. Restricted boltzmann machines for collaborative filtering. In Proceedings of the 24th international conference on Machine learning, pages 791–798, 2007

  34. [42]

    Statistical mechanics: A short treatise

    Giovanni Gallavotti. Statistical mechanics: A short treatise . Springer Science & Business Media, 1999

  35. [43]

    Flow contrastive estimation of energy-based models

    Ruiqi Gao, Erik Nijkamp, Diederik P Kingma, Zhen Xu, Andrew M Dai, and Ying Nian Wu. Flow contrastive estimation of energy-based models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7518–7528, 2020

  36. [44]

    On the anatomy of mcmc-based maximum likelihood learning of energy-based models

    Erik Nijkamp, Mitch Hill, Tian Han, Song-Chun Zhu, and Ying Nian Wu. On the anatomy of mcmc-based maximum likelihood learning of energy-based models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 5272–5280, 2020

  37. [45]

    Equation of state calculations by fast computing machines

    Nicholas Metropolis, Arianna W Rosenbluth, Marshall N Rosenbluth, Augusta H Teller, and Edward Teller. Equation of state calculations by fast computing machines. The journal of chemical physics, 21(6):1087–1092, 1953

  38. [46]

    Keith Hastings

    W. Keith Hastings. Monte carlo sampling methods using markov chains and their applications. Biometrika, 57(1):97–109, 1970

  39. [47]

    Stochastic relaxation, gibbs distributions, and the bayesian restoration of images

    Stuart Geman and Donald Geman. Stochastic relaxation, gibbs distributions, and the bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence , PAMI-6(6):721–741, 1984

  40. [48]

    Reflected diffusion models

    Aaron Lou and Stefano Ermon. Reflected diffusion models. In International Conference on Machine Learning, pages 22675–22701. PMLR, 2023

  41. [49]

    Flow++: Improving flow-based generative models with variational dequantization and architecture design

    Jonathan Ho, Xi Chen, Aravind Srinivas, Yan Duan, and Pieter Abbeel. Flow++: Improving flow-based generative models with variational dequantization and architecture design. In International conference on machine learning, pages 2722–2730. PMLR, 2019

  42. [50]

    Diffusion normalizing flow

    Qinsheng Zhang and Yongxin Chen. Diffusion normalizing flow. Advances in neural informa- tion processing systems, 34:16280–16291, 2021

  43. [51]

    Light and optimal schrödinger bridge matching

    Nikita Gushchin, Sergei Kholkin, Evgeny Burnaev, and Alexander Korotin. Light and optimal schrödinger bridge matching. In Forty-first International Conference on Machine Learning, 2024

  44. [52]

    Time, structure, and fluctuations

    Ilya Prigogine. Time, structure, and fluctuations. Science, 201(4358):777–785, 1978

  45. [53]

    Non-equilibrium statistical mechanics

    Ilya Prigogine. Non-equilibrium statistical mechanics. Courier Dover Publications, 2017

  46. [54]

    Equilibrium free-energy differences from nonequilibrium measure- ments: A master-equation approach

    Christopher Jarzynski. Equilibrium free-energy differences from nonequilibrium measure- ments: A master-equation approach. Physical Review E, 56(5):5018, 1997

  47. [55]

    Annealed importance sampling

    Radford M Neal. Annealed importance sampling. Statistics and computing, 11:125–139, 2001

  48. [56]

    Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences

    Gavin E Crooks. Entropy production fluctuation theorem and the nonequilibrium work relation for free energy differences. Physical Review E, 60(3):2721, 1999

  49. [57]

    Representations of knowledge in complex systems

    Ulf Grenander and Michael I Miller. Representations of knowledge in complex systems. Journal of the Royal Statistical Society: Series B (Methodological), 56(4):549–581, 1994

  50. [58]

    Stochastic differential equations

    Bernt Øksendal. Stochastic differential equations. Springer, 2003

  51. [59]

    Closing the ode-sde gap in score-based diffusion models through the fokker-planck equation

    Teo Deveney, Jan Stanczuk, Lisa Maria Kreusser, Chris Budd, and Carola-Bibiane Schönlieb. Closing the ode-sde gap in score-based diffusion models through the fokker-planck equation. arXiv preprint arXiv:2311.15996, 2023. 12

  52. [60]

    A connection between score matching and denoising autoencoders

    Pascal Vincent. A connection between score matching and denoising autoencoders. Neural computation, 23(7):1661–1674, 2011

  53. [61]

    Sliced score matching: A scalable approach to density and score estimation

    Yang Song, Sahaj Garg, Jiaxin Shi, and Stefano Ermon. Sliced score matching: A scalable approach to density and score estimation. In Uncertainty in Artificial Intelligence , pages 574–584. PMLR, 2020

  54. [62]

    Genphys: From physical processes to generative models

    Ziming Liu, Di Luo, Yilun Xu, Tommi Jaakkola, and Max Tegmark. Genphys: From physical processes to generative models. arXiv preprint arXiv:2304.02637, 2023

  55. [63]

    Flow matching for generative modeling

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling. arXiv preprint arXiv:2210.02747, 2022

  56. [64]

    A survey of the schr\" odinger problem and some of its connections with optimal transport

    Christian Léonard. A survey of the schr\" odinger problem and some of its connections with optimal transport. arXiv preprint arXiv:1308.0215, 2013

  57. [65]

    Optimal transport in systems and control

    Yongxin Chen, Tryphon T Georgiou, and Michele Pavon. Optimal transport in systems and control. Annual Review of Control, Robotics, and Autonomous Systems, 4(1):89–113, 2021

  58. [66]

    beta-vae: Learning basic visual concepts with a constrained variational framework

    Irina Higgins, Loic Matthey, Arka Pal, Christopher Burgess, Xavier Glorot, Matthew Botvinick, Shakir Mohamed, and Alexander Lerchner. beta-vae: Learning basic visual concepts with a constrained variational framework. In International conference on learning representations, 2017

  59. [67]

    An introduction to variational methods for graphical models

    Michael I Jordan, Zoubin Ghahramani, Tommi S Jaakkola, and Lawrence K Saul. An introduction to variational methods for graphical models. Machine learning, 37(2):183–233, 1999

  60. [68]

    Information theory, inference and learning algorithms

    David JC MacKay. Information theory, inference and learning algorithms . Cambridge university press, 2003

  61. [69]

    Wavenet: A generative model for raw audio

    Aaron van den Oord, Sander Dieleman, Heiga Zen, et al. Wavenet: A generative model for raw audio. arXiv preprint arXiv:1609.03499, 2016

  62. [70]

    Generative pretraining from pixels

    Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, and Ilya Sutskever. Generative pretraining from pixels. In International conference on machine learning, pages 1691–1703. PMLR, 2020

  63. [71]

    A neural probabilistic language model

    Yoshua Bengio, Réjean Ducharme, Pascal Vincent, and Christian Jauvin. A neural probabilistic language model. Journal of Machine Learning Research, 3:1137–1155, 2003

  64. [72]

    Generating text with recurrent neural networks

    Ilya Sutskever, James Martens, and Geoffrey E Hinton. Generating text with recurrent neural networks. In Proceedings of the 28th international conference on machine learning (ICML-11), pages 1017–1024, 2011

  65. [73]

    Language models are few-shot learners

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhari- wal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877– 1901, 2020

  66. [74]

    Magvit: Masked generative video transformer

    Lijun Yu, Yong Cheng, Kihyuk Sohn, José Lezama, Han Zhang, Huiwen Chang, Alexander G Hauptmann, Ming-Hsuan Yang, Yuan Hao, Irfan Essa, et al. Magvit: Masked generative video transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, page...

  67. [75]

    Unsupervised representation learning with deep convolutional generative adversarial networks

    Alec Radford, Luke Metz, and Soumith Chintala. Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv preprint arXiv:1511.06434, 2015

  68. [76]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock, Jeff Donahue, and Karen Simonyan. Large scale gan training for high fidelity natural image synthesis. arXiv preprint arXiv:1809.11096, 2018

  69. [77]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4401–4410, 2019. 13

  70. [78]

    Analyzing and improving the image quality of stylegan

    Tero Karras, Samuli Laine, Miika Aittala, Janne Hellsten, Jaakko Lehtinen, and Timo Aila. Analyzing and improving the image quality of stylegan. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8110–8119, 2020

  71. [79]

    Predicting the energy landscape of stochastic dynamical system via physics-informed self-supervised learning

    Ruikun Li, Huandong Wang, Qingmin Liao, and Yong Li. Predicting the energy landscape of stochastic dynamical system via physics-informed self-supervised learning. arXiv preprint arXiv:2502.16828, 2025

  72. [80]

    Latent representation and simulation of markov processes via time-lagged information bottleneck

    Marco Federici, Patrick Forré, Ryota Tomioka, and Bastiaan S Veeling. Latent representation and simulation of markov processes via time-lagged information bottleneck. arXiv preprint arXiv:2309.07200, 2023

  73. [81]

    Emergence of polarization in coevolving networks

    Jiazhen Liu, Shengda Huang, Nathaniel M Aden, Neil F Johnson, and Chaoming Song. Emergence of polarization in coevolving networks. Physical Review Letters, 130(3):037401, 2023

  74. [82]

    Correlated growth of causal networks

    Jiazhen Liu, Kunal Tamang, Dashun Wang, and Chaoming Song. Correlated growth of causal networks. arXiv preprint arXiv:2412.16647, 2024

  75. [83]

    Dynamical phase transitions in non-equilibrium networks

    Jiazhen Liu, Nathaniel M Aden, Debasish Sarker, and Chaoming Song. Dynamical phase transitions in non-equilibrium networks. arXiv preprint arXiv:2412.06704, 2024

  76. [84]

    Non-equilibrium statistical mechanics: from a paradigmatic model to biological transport

    Tom Chou, Kirone Mallick, and Royce KP Zia. Non-equilibrium statistical mechanics: from a paradigmatic model to biological transport. Reports on Progress in Physics, 74(11):116601,

  77. [85]

    Brownian motors: noisy transport far from equilibrium

    Peter Reimann. Brownian motors: noisy transport far from equilibrium. Physics Reports, 361 (2-4):57–265, 2002. doi: 10.1016/S0370-1573(01)00081-3

  78. [86]

    Learning nonequilibrium statistical mechanics and dynamical phase transitions

    Ying Tang, Jing Liu, Jiang Zhang, and Pan Zhang. Learning nonequilibrium statistical mechanics and dynamical phase transitions. Nature Communications, 15(1):1117, 2024

  79. [87]

    Ai-nerd: Elucidation of relaxation dynamics beyond equilibrium via deep learning

    Huili Liao, Xiaoyu Huang, Siyu Xiang, et al. Ai-nerd: Elucidation of relaxation dynamics beyond equilibrium via deep learning. Nature Communications, 15(1):4938, 2024. doi: 10.1038/s41467-024-45667-7

  80. [88]

    Stochastic optimal control for collective variable free sampling of molecular transition paths

    Lars Holdijk, Yuanqi Du, Ferry Hooft, Priyank Jaini, Berend Ensing, and Max Welling. Stochastic optimal control for collective variable free sampling of molecular transition paths. Advances in Neural Information Processing Systems, 36:79540–79556, 2023

  81. [89]

    Flowmm: Generating materials with riemannian flow matching

    Benjamin Kurt Miller, Ricky TQ Chen, Anuroop Sriram, and Brandon M Wood. Flowmm: Generating materials with riemannian flow matching. In Forty-first International Conference on Machine Learning, 2024

  82. [90]

    Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties

    Tian Xie and Jeffrey C Grossman. Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Physical review letters, 120(14):145301, 2018

  83. [91]

    Crystal structure prediction by joint equivariant diffusion

    Rui Jiao, Wenbing Huang, Peijia Lin, Jiaqi Han, Pin Chen, Yutong Lu, and Yang Liu. Crystal structure prediction by joint equivariant diffusion. Advances in Neural Information Processing Systems, 36:17464–17497, 2023

  84. [92]

    Molgan: An implicit generative model for small molecular graphs

    Nicola De Cao and Thomas Kipf. Molgan: An implicit generative model for small molecular graphs. arXiv preprint arXiv:1805.11973, 2018

  85. [93]

    Diffbp: Generative diffusion of 3d molecules for target protein binding

    Haitao Lin, Yufei Huang, Odin Zhang, Siqi Ma, Meng Liu, Xuanjing Li, Lirong Wu, Jishui Wang, Tingjun Hou, and Stan Z Li. Diffbp: Generative diffusion of 3d molecules for target protein binding. Chemical Science, 16(3):1417–1431, 2025

  86. [94]

    Geod- iff: A geometric diffusion model for molecular conformation generation

    Minkai Xu, Lantao Yu, Yang Song, Chence Shi, Stefano Ermon, and Jian Tang. Geod- iff: A geometric diffusion model for molecular conformation generation. arXiv preprint arXiv:2203.02923, 2022. 14

  87. [95]

    A generative model for inorganic materials design

    Claudio Zeni, Robert Pinsler, Daniel Zügner, Andrew Fowler, Matthew Horton, Xiang Fu, Zilong Wang, Aliaksandra Shysheya, Jonathan Crabbé, Shoko Ueda, et al. A generative model for inorganic materials design. Nature, pages 1–3, 2025

  88. [96]

    Symmcd: Symmetry-preserving crystal generation with diffusion models

    Daniel Levy, Siba Smarak Panigrahi, Sékou-Oumar Kaba, Qiang Zhu, Kin Long Kelvin Lee, Mikhail Galkin, Santiago Miret, and Siamak Ravanbakhsh. Symmcd: Symmetry-preserving crystal generation with diffusion models. arXiv preprint arXiv:2502.03638, 2025

  89. [97]

    Model scale versus domain knowledge in statistical forecasting of chaotic systems

    William Gilpin. Model scale versus domain knowledge in statistical forecasting of chaotic systems. Physical Review Research, 5(4):043252, 2023

  90. [98]

    Predicting long-term dynamics of complex networks via identifying skeleton in hyperbolic space

    Ruikun Li, Huandong Wang, Jinghua Piao, Qingmin Liao, and Yong Li. Predicting long-term dynamics of complex networks via identifying skeleton in hyperbolic space. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages 1655–1666, 2024

  91. [99]

    Integrating artificial intelligence with mechanistic epidemiological modeling: a scoping review of opportunities and challenges

    Yang Ye, Abhishek Pandey, Carolyn Bawden, Dewan Md Sumsuzzman, Rimpi Rajput, Affan Shoukat, Burton H Singer, Seyed M Moghadas, and Alison P Galvani. Integrating artificial intelligence with mechanistic epidemiological modeling: a scoping review of opportunities and challenges....

  92. [100]

    Artificial intelligence for complex network: Potential, methodology and application

    Jingtao Ding, Chang Liu, Yu Zheng, Yunke Zhang, Zihan Yu, Ruikun Li, Hongyi Chen, Jinghua Piao, Huandong Wang, Jiazhen Liu, et al. Artificial intelligence for complex network: Potential, methodology and application. arXiv preprint arXiv:2402.16887, 2024

  93. [101]

    Graphite: Iterative generative modeling of graphs

    Aditya Grover, Aaron Zweig, and Stefano Ermon. Graphite: Iterative generative modeling of graphs. In International conference on machine learning, pages 2434–2444. PMLR, 2019

  94. [102]

    Discovering governing equations from data by sparse identification of nonlinear dynamical systems.Proceedings of the National Academy of Sciences, 113(15):3932–3937, 2016

    Steven L Brunton, Joshua L Proctor, and J Nathan Kutz. Discovering governing equations from data by sparse identification of nonlinear dynamical systems.Proceedings of the National Academy of Sciences, 113(15):3932–3937, 2016

  95. [103]

    Coupling functions: Universal insights into dynamical interaction mechanisms

    Tomislav Stankovski, Tiago Pereira, Peter VE McClintock, and Aneta Stefanovska. Coupling functions: Universal insights into dynamical interaction mechanisms. Reviews of Modern Physics, 89(4):045001, 2017

  96. [104]

    Autonomous inference of complex network dynamics from incomplete and noisy data

    Tian-Tian Gao and Gang Yan. Autonomous inference of complex network dynamics from incomplete and noisy data. Nature Computational Science, 2:160–168, 2022

  97. [105]

    Causal inference in statistics, social, and biomedical sciences

    Guido W Imbens and Donald B Rubin. Causal inference in statistics, social, and biomedical sciences. Cambridge university press, 2015

  98. [106]

    Causality: Models, Reasoning and Inference

    Judea Pearl. Causality: Models, Reasoning and Inference . Cambridge University Press, Cambridge, 2000

  99. [107]

    Estimating causal effects of treatments in randomized and nonrandomized studies

    Donald B Rubin. Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of Educational Psychology, 66(5):688–701, 1974. doi: 10.1037/h0037350

  100. [108]

    Seneviratne, Thomas R

    Jakob Zscheischler, Sonia I. Seneviratne, Thomas R. Karl, and et al. Understanding and managing connected extreme events. Nature Climate Change , 10:611–621, 2020. doi: 10.1038/s41558-020-0790-4

  101. [109]

    Critical slowing down at the abrupt mott transition: When the first-order phase transition becomes zeroth- order and looks like second-order

    Satyaki Kundu, Tapas Bar, Rajesh Kumble Nayak, and Bhavtosh Bansal. Critical slowing down at the abrupt mott transition: When the first-order phase transition becomes zeroth- order and looks like second-order. Physical Review Letters , 124(9):095703, 2020. doi: 10.1103/PhysRev...

  102. [110]

    Microscopic origin of abrupt mixed- order phase transitions

    Jan Korbel, Shlomo Havlin, and Stefan Thurner. Microscopic origin of abrupt mixed- order phase transitions. Nature Communications , 16(1):2628, 2025. doi: 10.1038/ s41467-025-57007-1

  103. [111]

    Almendral, Shuguang Guan, Iacopo Leyva, Zonghua Liu, Iker Sendiña-Nadal, Zhen Wang, and Yamir Moreno

    Stefano Boccaletti, Juan A. Almendral, Shuguang Guan, Iacopo Leyva, Zonghua Liu, Iker Sendiña-Nadal, Zhen Wang, and Yamir Moreno. Explosive transitions in complex networks’ structure and dynamics: percolation and synchronization. Physics Reports, 660:1–94, 2016. doi: 10.1016/j...

  104. [112]

    Probability of second law violations in shearing steady states

    Denis J Evans, Ezechiel Godert David Cohen, and Gary P Morriss. Probability of second law violations in shearing steady states. Physical review letters, 71(15):2401, 1993

  105. [113]

    Dynamical ensembles in nonequilib- rium statistical mechanics

    Giovanni Gallavotti and Ezechiel Godert David Cohen. Dynamical ensembles in nonequilib- rium statistical mechanics. Physical review letters, 74(14):2694, 1995

  106. [114]

    Dynamicsdiffusion: Generating and rare event sampling of molecular dynamic trajectories using diffusion models

    Magnus Petersen, Gemma Roig, and Roberto Covino. Dynamicsdiffusion: Generating and rare event sampling of molecular dynamic trajectories using diffusion models

  107. [115]

    Generating images of rare concepts using pre-trained diffusion models

    Dvir Samuel, Rami Ben-Ari, Simon Raviv, Nir Darshan, and Gal Chechik. Generating images of rare concepts using pre-trained diffusion models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 4695–4703, 2024

  108. [116]

    Multi-scale simulation of complex systems: a perspective of integrating knowledge and data

    Huandong Wang, Huan Yan, Can Rong, Yuan Yuan, Fenyu Jiang, Zhenyu Han, Hongjie Sui, Depeng Jin, and Yong Li. Multi-scale simulation of complex systems: a perspective of integrating knowledge and data. ACM Computing Surveys, 56(12):1–38, 2024

  109. [117]

    Physical and scale-by-scale analysis of rayleigh–bénard convection

    Riccardo Togni, Andrea Cimarelli, and Elisabetta De Angelis. Physical and scale-by-scale analysis of rayleigh–bénard convection. Journal of Fluid Mechanics, 782:380–404, 2015

  110. [118]

    Particle-fluid two-phase flow: the energy-minimization multi-scale method

    Jinghai Li. Particle-fluid two-phase flow: the energy-minimization multi-scale method. Metal- lurgical Industry Press, 1994

  111. [119]

    Physical mapping of fluidization regimes—the emms approach

    Wei Ge and Jinghai Li. Physical mapping of fluidization regimes—the emms approach. Chemical Engineering Science, 57(18):3993–4004, 2002

  112. [120]

    Design of multi-scale protein complexes by hierarchical building block fusion

    Yang Hsia, Rubul Mout, William Sheffler, Natasha I Edman, Ivan Vulovic, Young-Jun Park, Rachel L Redler, Matthew J Bick, Asim K Bera, Alexis Courbet, et al. Design of multi-scale protein complexes by hierarchical building block fusion. Nature Communications, 12(1):2294, 2021

  113. [121]

    Mg-tsd: Multi-granularity time series diffusion models with guided learning process

    Xinyao Fan, Yueying Wu, Chang Xu, Yuhao Huang, Weiqing Liu, and Jiang Bian. Mg-tsd: Multi-granularity time series diffusion models with guided learning process. In The Twelfth International Conference on Learning Representations, 2024

  114. [122]

    Generative emulation of weather forecast ensembles with diffusion models

    Lizao Li, Robert Carver, Ignacio Lopez-Gomez, Fei Sha, and John Anderson. Generative emulation of weather forecast ensembles with diffusion models. Science Advances, 10(13): eadk4489, 2024

  115. [123]

    Condi- tional neural field latent diffusion model for generating spatiotemporal turbulence

    Pan Du, Meet Hemant Parikh, Xiantao Fan, Xin-Yang Liu, and Jian-Xun Wang. Condi- tional neural field latent diffusion model for generating spatiotemporal turbulence. Nature Communications, 15(1):10416, 2024

  116. [124]

    Generative diffusion model with inverse renormalization group flows

    Kanta Masuki and Yuto Ashida. Generative diffusion model with inverse renormalization group flows. arXiv preprint arXiv:2501.09064, 2025

  117. [125]

    Quantum state generation with structure-preserving diffusion model

    Yuchen Zhu, Tianrong Chen, Evangelos A Theodorou, Xie Chen, and Molei Tao. Quantum state generation with structure-preserving diffusion model. arXiv preprint arXiv:2404.06336, 2024

  118. [126]

    Coupled graph ode for learning interacting system dynamics

    Zijie Huang, Yizhou Sun, and Wei Wang. Coupled graph ode for learning interacting system dynamics. In Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining, pages 705–715, 2021

  119. [127]

    Protein structure generation via folding diffusion

    Kevin E Wu, Kevin K Yang, Rianne van den Berg, Sarah Alamdari, James Y Zou, Alex X Lu, and Ava P Amini. Protein structure generation via folding diffusion. Nature communications, 15(1):1059, 2024

  120. [128]

    Multi-scale representation learning on proteins

    Vignesh Ram Somnath, Charlotte Bunne, and Andreas Krause. Multi-scale representation learning on proteins. Advances in Neural Information Processing Systems, 34:25244–25255, 2021

  121. [129]

    A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect

    James M Robins. A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect. Mathematical Modelling, 7(9–12):1393–1512, 1986. doi: 10.1016/0270-0255(86)90088-6. 16

  122. [130]

    Causal inference using graphical models with the r package pcalg

    Bernhard Schölkopf et al. Causal inference using graphical models with the r package pcalg. In Proceedings of the 29th International Conference on Machine Learning (ICML), 2012

  123. [131]

    Elements of Causal Inference: Foundations and Learning Algorithms

    Jonas Peters, Dominik Janzing, and Bernhard Schölkopf. Elements of Causal Inference: Foundations and Learning Algorithms. MIT Press, 2017

  124. [132]

    A meta-transfer objective for learning to disentangle causal mechanisms

    Yoshua Bengio, Tristan Deleu, Nasim Rahaman, Nan Ke, Sebastien Lachapelle, Olexa Bilaniuk, Anirudh Goyal, and Chris Pal. A meta-transfer objective for learning to disentangle causal mechanisms. In Advances in Neural Information Processing Systems (NeurIPS), volume 33, 2020

  125. [133]

    Tensor field networks: Rotation-and translation-equivariant neural networks for 3d point clouds

    Nathaniel Thomas, Tess Smidt, Steven Kearnes, Lusann Yang, Li Li, Kai Kohlhoff, and Patrick Riley. Tensor field networks: Rotation-and translation-equivariant neural networks for 3d point clouds. arXiv preprint arXiv:1802.08219, 2018

  126. [134]

    Hamiltonian neural networks

    Samuel Greydanus, Misko Dzamba, and Jason Yosinski. Hamiltonian neural networks. Ad- vances in neural information processing systems, 32, 2019

  127. [135]

    E (3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials

    Simon Batzner, Albert Musaelian, Lixin Sun, Mario Geiger, Jonathan P Mailoa, Mordechai Kornbluth, Nicola Molinari, Tess E Smidt, and Boris Kozinsky. E (3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nature communications, 13(1): 2453, 2022

  128. [136]

    Noether’s learning dynamics: Role of symmetry breaking in neural networks

    Hidenori Tanaka and Daniel Kunin. Noether’s learning dynamics: Role of symmetry breaking in neural networks. Advances in Neural Information Processing Systems, 34:25646–25660, 2021

  129. [137]

    Group equivariant convolutional networks

    Taco Cohen and Max Welling. Group equivariant convolutional networks. In International conference on machine learning, pages 2990–2999. PMLR, 2016

  130. [138]

    On the generalization of equivariance and convolution in neural networks to the action of compact groups

    Risi Kondor and Shubhendu Trivedi. On the generalization of equivariance and convolution in neural networks to the action of compact groups. In International conference on machine learning, pages 2747–2755. PMLR, 2018

  131. [139]

    Generalizing convolutional neural networks for equivariance to lie groups on arbitrary continuous data

    Marc Finzi, Samuel Stanton, Pavel Izmailov, and Andrew Gordon Wilson. Generalizing convolutional neural networks for equivariance to lie groups on arbitrary continuous data. In International Conference on Machine Learning, pages 3165–3176. PMLR, 2020

  132. [140]

    Model-free predic- tion of large spatiotemporally chaotic systems from data: A reservoir computing approach

    Jaideep Pathak, Brian Hunt, Michelle Girvan, Zhixin Lu, and Edward Ott. Model-free predic- tion of large spatiotemporally chaotic systems from data: A reservoir computing approach. Physical review letters, 120(2):024102, 2018

  133. [141]

    Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning

    Frank Noé, Simon Olsson, Jonas Köhler, and Hao Wu. Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning. Science, 365(6457):eaaw1147, 2019

  134. [142]

    The schrödinger bridge between gaussian measures has a closed form

    Charlotte Bunne, Ya-Ping Hsieh, Marco Cuturi, and Andreas Krause. The schrödinger bridge between gaussian measures has a closed form. In Proceedings of the 40th International Conference on Machine Learning, pages 1234–1243, 2023

  135. [143]

    Computational optimal transport: With applications to data science

    Gabriel Peyré and Marco Cuturi. Computational optimal transport: With applications to data science. Foundations and Trends in Machine Learning, 11(5-6):355–607, 2019

  136. [144]

    Predicting the dynamics of complex system via multiscale diffusion autoencoder

    Ruikun Li, Jingwen Cheng, Huandong Wang, Qingmin Liao, and Yong Li. Predicting the dynamics of complex system via multiscale diffusion autoencoder. arXiv preprint arXiv:2505.02450, 2025

  137. [145]

    Interpretable scientific discovery with symbolic regression: a review

    Nour Makke and Sanjay Chawla. Interpretable scientific discovery with symbolic regression: a review. Artificial Intelligence Review, 57(1):2, 2024

  138. [146]

    Symbolic regression via mdlformer-guided search: from minimizing prediction error to minimizing description length

    Zihan Yu, Jingtao Ding, and Yong Li. Symbolic regression via mdlformer-guided search: from minimizing prediction error to minimizing description length. In The Thirteenth International Conference on Learning Representations, 2025. 17

  139. [147]

    Learning symbolic models for graph-structured physical mechanism

    Hongzhi Shi, Jingtao Ding, Yufan Cao, Li Liu, Yong Li, et al. Learning symbolic models for graph-structured physical mechanism. In The Eleventh International Conference on Learning Representations, 2023

  140. [148]

    Neural sdes as a unified approach to continuous-domain sequence modeling

    Macheng Shen and Chen Cheng. Neural sdes as a unified approach to continuous-domain sequence modeling. arXiv preprint arXiv:2501.18871, 2025

  141. [149]

    Machine learning for advancing low- temperature plasma modeling and simulation

    Jan Trieschmann, Luca Vialetto, and Tobias Gergs. Machine learning for advancing low- temperature plasma modeling and simulation. Journal of Micro/Nanopatterning, Materials, and Metrology, 22(4):041504–041504, 2023

  142. [150]

    Solving the quantum many-body problem with artificial neural networks

    Giuseppe Carleo and Matthias Troyer. Solving the quantum many-body problem with artificial neural networks. Science, 355(6325):602–606, 2017

  143. [151]

    Information-theoretic diffusion

    Xianghao Kong, Rob Brekelmans, and Greg Ver Steeg. Information-theoretic diffusion. In The Eleventh International Conference on Learning Representations

  144. [152]

    Deep learning probability flows and entropy production rates in active matter

    Nicholas M Boffi and Eric Vanden-Eijnden. Deep learning probability flows and entropy production rates in active matter. Proceedings of the National Academy of Sciences, 121(25): e2318106121, 2024

  145. [153]

    Optimal transport: old and new, volume 338

    Cédric Villani et al. Optimal transport: old and new, volume 338. Springer, 2009

  146. [154]

    The fluctuation-dissipation theorem

    Rep Kubo. The fluctuation-dissipation theorem. Reports on progress in physics, 29(1):255, 1966

  147. [155]

    Neural variable-order fractional differential equation networks

    Wenjun Cui, Qiyu Kang, Xuhao Li, Kai Zhao, Wee Peng Tay, Weihua Deng, and Yidong Li. Neural variable-order fractional differential equation networks. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 16109–16117, 2025

  148. [156]

    Machine learning non- markovian quantum dynamics

    IA Luchnikov, SV Vintskevich, DA Grigoriev, and SN Filippov. Machine learning non- markovian quantum dynamics. Physical review letters, 124(14):140502, 2020

  149. [157]

    Neural network aided approximation and parameter inference of non-markovian models of gene expression

    Qingchao Jiang, Xiaoming Fu, Shifu Yan, Runlai Li, Wenli Du, Zhixing Cao, Feng Qian, and Ramon Grima. Neural network aided approximation and parameter inference of non-markovian models of gene expression. Nature communications, 12(1):2618, 2021

  150. [158]

    Learning slow and fast system dynamics via automatic separation of time scales

    Ruikun Li, Huandong Wang, and Yong Li. Learning slow and fast system dynamics via automatic separation of time scales. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 4380–4390, 2023

  151. [159]

    Learning gradient fields for molecular conformation generation

    Chence Shi, Shitong Luo, Minkai Xu, and Jian Tang. Learning gradient fields for molecular conformation generation. In ICML, pages 9558–9568. PMLR, 2021

  152. [160]

    Predicting molecular conformation via dynamic graph score matching

    Shitong Luo, Chence Shi, Minkai Xu, and Jian Tang. Predicting molecular conformation via dynamic graph score matching. NeurIPS, 34:19784–19795, 2021

  153. [161]

    Torsional diffusion for molecular conformer generation

    Bowen Jing, Gabriele Corso, Jeffrey Chang, Regina Barzilay, and Tommi Jaakkola. Torsional diffusion for molecular conformer generation. Advances in Neural Information Processing Systems, 35:24240–24253, 2022

  154. [162]

    Protein structure and sequence generation with equivariant denoising diffusion probabilistic models

    Namrata Anand and Tudor Achim. Protein structure and sequence generation with equivariant denoising diffusion probabilistic models. arXiv preprint arXiv:2205.15019, 2022

  155. [163]

    Antigen- specific antibody design and optimization with diffusion-based generative models for protein structures

    Shitong Luo, Yufeng Su, Xingang Peng, Sheng Wang, Jian Peng, and Jianzhu Ma. Antigen- specific antibody design and optimization with diffusion-based generative models for protein structures. Advances in Neural Information Processing Systems, 35:9754–9767, 2022

  156. [164]

    De novo design of protein structure and function with rfdiffusion

    Joseph L Watson, David Juergens, Nathaniel R Bennett, Brian L Trippe, Jason Yim, Helen E Eisenach, Woody Ahern, Andrew J Borst, Robert J Ragotte, Lukas F Milles, et al. De novo design of protein structure and function with rfdiffusion. Nature, 620(7976):1089–1100, 2023. 18

  157. [165]

    Illumi- nating protein space with a programmable generative model

    John B Ingraham, Max Baranov, Zak Costello, Karl W Barber, Wujie Wang, Ahmed Ismail, Vincent Frappier, Dana M Lord, Christopher Ng-Thow-Hing, Erik R Van Vlack, et al. Illumi- nating protein space with a programmable generative model. Nature, 623(7989):1070–1078, 2023

  158. [166]

    Score-based generative modeling for de novo protein design

    Jin Sub Lee, Jisun Kim, and Philip M Kim. Score-based generative modeling for de novo protein design. Nature Computational Science, 3(5):382–392, 2023

  159. [167]

    Mr image denoising and super-resolution using regularized reverse diffusion

    Hyungjin Chung, Eun Sun Lee, and Jong Chul Ye. Mr image denoising and super-resolution using regularized reverse diffusion. IEEE Transactions on Medical Imaging, 42(4):922–934, 2022

  160. [168]

    Solving inverse problems in medical imaging with score-based generative models

    Yang Song, Liyue Shen, Lei Xing, and Stefano Ermon. Solving inverse problems in medical imaging with score-based generative models. ICLR, 2022

  161. [169]

    Robust compressed sensing mri with deep generative priors

    Ajil Jalal, Marius Arvinte, Giannis Daras, Eric Price, Alexandros G Dimakis, and Jon Tamir. Robust compressed sensing mri with deep generative priors. NeurIPS, 34:14938–14954, 2021

  162. [170]

    Score-based diffusion models for accelerated mri.Medical image analysis, 80:102479, 2022

    Hyungjin Chung and Jong Chul Ye. Score-based diffusion models for accelerated mri.Medical image analysis, 80:102479, 2022

  163. [171]

    High-frequency space diffusion model for accelerated mri

    Chentao Cao, Zhuo-Xu Cui, Yue Wang, Shaonan Liu, Taijin Chen, Hairong Zheng, Dong Liang, and Yanjie Zhu. High-frequency space diffusion model for accelerated mri. IEEE Transactions on Medical Imaging, 2024

  164. [172]

    Towards performant and reliable undersampled mr reconstruction via diffusion model sampling

    Cheng Peng, Pengfei Guo, S Kevin Zhou, Vishal M Patel, and Rama Chellappa. Towards performant and reliable undersampled mr reconstruction via diffusion model sampling. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 623–633. Spr...

  165. [173]

    Measurement-conditioned denoising diffusion probabilistic model for under-sampled medical image reconstruction

    Yutong Xie and Quanzheng Li. Measurement-conditioned denoising diffusion probabilistic model for under-sampled medical image reconstruction. In International Conference on Medical Image Computing and Computer-Assisted Intervention , pages 655–664. Springer, 2022

  166. [174]

    A physics-informed diffusion model for high-fidelity flow field reconstruction

    Dule Shu, Zijie Li, and Amir Barati Farimani. A physics-informed diffusion model for high-fidelity flow field reconstruction. Journal of Computational Physics, 478:111972, 2023

  167. [175]

    Generative prediction of flow field based on the diffusion model

    Jiajun Hu, Zhen Lu, and Yue Yang. Generative prediction of flow field based on the diffusion model. arXiv preprint arXiv:2407.00735, 2024

  168. [176]

    Probabilistic weather forecasting with machine learning

    Ilan Price, Alvaro Sanchez-Gonzalez, Ferran Alet, Tom R Andersson, Andrew El-Kadi, Dominic Masters, Timo Ewalds, Jacklynn Stott, Shakir Mohamed, Peter Battaglia, et al. Probabilistic weather forecasting with machine learning. Nature, 637(8044):84–90, 2025

  169. [177]

    Radio- astronomical image reconstruction with a conditional denoising diffusion model

    Mariia Drozdova, Vitaliy Kinakh, Omkar Bait, Olga Taran, Erica Lastufka, Miroslava Dessauges-Zavadsky, Taras Holotyak, Daniel Schaerer, and Slava V oloshynovskiy. Radio- astronomical image reconstruction with a conditional denoising diffusion model. Astronomy & Astrophysics, 6...

  170. [178]

    free energy

    Renato Sortino, Thomas Cecconello, Andrea DeMarco, Giuseppe Fiameni, Andrea Pilzer, Daniel Magro, Andrew M Hopkins, Simone Riggi, Eva Sciacca, Adriano Ingallinera, et al. Rad- iff: controllable diffusion models for radio astronomical maps generation. IEEE Transactions on Artif...

  171. [180]

    [178] propose using the diffusion model for astronomical image denoising and synthesis, showing more efficient and effective performance than other methods

    and Sortino et al. [178] propose using the diffusion model for astronomical image denoising and synthesis, showing more efficient and effective performance than other methods. 26

  172. [2011]

    doi: 10.1088/0034-4885/74/11/116601

Pith tools

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