Pith. sign in

REVIEW 4 major objections 4 minor 28 references

Hybrid Generative Modeling for Incomplete Physics: Deep Grey-Box Meets Optimal Transport

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper proposes completing an imperfect physics model from unpaired data by training a conditional optimal-transport map inside a deep grey-box model, and shows it outperforms WGAN and black-box baselines on joint…

desk verdict A useful combination of weak OT and grey-box neural ODEs for unpaired physics completion, but the conditional-correctness claim outruns the method and there's a missing cross-term in the stated cost. read the letter →

arxiv 2506.22204 v1 pith:GKRQG5ZQ submitted 2025-06-27 stat.ML cs.LG

Gurjeet Sangra Singh , Maciej Falkiewicz , Alexandros Kalousis This is my paper · ORCID
classification stat.MLcs.LG
keywords optimaltransportgrey-boxmodelinghybridgenerativemodelsincompletephysicsunpairedtranslationneuralODEsweakphysics-informedmachinelearning
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 paper tackles a common failure mode in scientific modeling: the governing ODE or PDE is known but incomplete, so trajectories simulated from it (source samples) have a different distribution than real observations (target samples), and no pairing between the two is available. Its central proposal is to learn a conditional map, built as a deep grey-box model that composes a learned neural component with the known physics block, and to train that map through a weak optimal-transport objective so that the pushforward of the simulator distribution matches the observed distribution with minimal distortion. The argument is that optimal transport's least-action cost keeps the correction minimal, while the grey-box composition keeps the physics block in the loop so that physics parameters $\theta$ keep their intended meaning. On damped pendulum, reaction-diffusion, and advection-diffusion systems, the OT grey-box variant beats Wasserstein-GAN and black-box baselines on joint metrics over input, parameter, and output—suggesting the method produces both accurate and interpretable completions.

What carries the argument

The central objects are (i) the conditional grey-box map $T_\varphi(x; \theta, z) = \mathrm{ODESolve}(dy/dt = f_\varphi(y(t), \theta, z) \circ f_p(y(t), \theta))$, a Neural ODE in which the learned network composes the physics block, and (ii) the weak optimal-transport objective (Equation 3) with kernel weak quadratic cost $C_{k,\gamma}$ (Equation 4), solved adversarially by alternating a generator/potential pair with gradient penalty. The weak-OT formulation allows one-to-many stochastic maps through the latent $z$ and $\gamma = 1$, while the grey-box composition acts as a soft constraint that forces the learned component to act through the physics model.

What would settle it

Take a trained OT grey-box model on a deterministic task (e.g., the one-to-one pendulum) and condition it on two clearly different physics parameters, such as ω = 0.8 and ω = 3.0, while holding the initial condition fixed; the true DGP produces trajectories with very different oscillation frequencies. If the generated trajectories do not show the corresponding frequency differences, or if the outputs are statistically indistinguishable for parameters that produce distinct true dynamics, then the claim of correct physics-parameter usage is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that a conditional stochastic map $T_\varphi(x; \theta, z)$, realized by a Neural ODE whose vector field composes a learned deep network $f_\varphi$ with the known (incomplete) physics model $f_p$, can implement the optimal transport between the simulator's source distribution and the data-generating distribution while maintaining minimal source-distortion. The optimal-transport objective—a weak-OT maximin problem with a kernel-based weak quadratic cost—aligns the marginal distributions in data space, and the grey-box architecture ensures the physics block $f_p$ continues to process the parameters $\theta$, so the learned component complements rather than overrides the known dynamics. On one-to-one and one-to-many translation tasks, the paper reports that this OT grey-box model outperforms WGAN-based grey-box models and purely data-driven black-box models when evaluated on the true joint distribution $\pi(x, \theta, y)$, and that component analysis reveals smooth, physically meaningful learned corrections.

Load-bearing premise

The model is trained only to match the marginal distribution of target trajectories, conditioned on a map tied to the physics block; the paper assumes this also produces the correct joint association between each input (x, θ) and its output y—that the learned map uses θ correctly.

Editorial extensions

If this is right

  • With a fixed simulation budget and unpaired target trajectories, the OT grey-box map can generate trajectories that match the target distribution more closely than WGAN-based or black-box alternatives, particularly for stochastic (one-to-many) dynamics.
  • The grey-box composition keeps the known physics block $f_p$ acting on the parameters $\theta$, so the learned correction stays subordinate; component analysis of $f_\varphi$ can expose the functional form of the missing physics.
  • The weak-OT cost with $\gamma=1$ and latent $z$ gives a principled way to learn multiple plausible continuations per input, which is the right inductive bias for stochastic physical systems.
  • Because marginal alignment alone does not guarantee correct conditional parameter usage, the paper argues that evaluation must be done on the joint distribution $\pi(x,\theta,y)$; its proposed metrics (N-RMSE/ABS for deterministic, MMD/C2ST for stochastic) are a template for that evaluation.

Reading between the lines

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

  • If the method works as claimed, a natural next step is to use the learned component's functional form as a hypothesis generator for the missing physics term, then add that term back into $f_p$ and re-run the same OT grey-box pipeline—this would turn the model from a correction into a discovery tool.
  • The marginal-matching assumption is the fragile point. For misspecifications where the missing term's effect cancels in the marginal distribution, or where $f_p$ is so flexible that the network could memorize spurious associations, the trained map could match target marginals while still misusing $\theta$; a stress test would condition the trained map on counterfactual parameter pairs to see wheth
  • The 'minimal source distortion' claim is tied to the choice of ground cost; using a different kernel or a different weak cost would likely change which completion is selected. The paper does not test how sensitive the learned correction is to that choice, so a robustness study across costs would clarify how much of the success is OT versus the grey-box architecture.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The manuscript proposes a hybrid generative model for correcting incomplete physics-based ODE/PDE simulators in an unpaired setting. It trains a conditional map Tφ(x; θ, z) by minimizing a weak-OT divergence between the pushforward of the source simulator distribution and the target DGP marginal, where the map is constrained by a grey-box composition fφ ∘ fp. The authors compare OT and WGAN variants, and grey-box versus black-box architectures, on three synthetic tasks (damped pendulum, reaction-diffusion, advection-diffusion), using joint metrics (N-RMSE/ABS for deterministic tasks, MMD and C2ST for stochastic tasks).

Significance. The paper addresses a relevant problem: completing misspecified physics models from unpaired simulations and observations. The weak-OT formulation gives a principled way to handle one-to-many stochastic mappings, and the grey-box component offers a path toward interpretability. The authors provide error bars, held-out joint evaluation, and a public code repository, which are commendable. However, the central claim that the method 'ensures correct usage of physics parameters' is not backed by the training objective or by an identifiability result; the paper itself acknowledges the marginal-alignment caveat but responds only with evaluation. The empirical joint scores leave room for conditional misspecification, so the significance is currently below what the abstract promises.

major comments (4)
  1. [Section 2, Eq. (3) and Algorithm C] The training objective is a weak-OT divergence between the pushforward marginal Tφ#(μ⊗η) and ν(y); there is no term comparing Tφ(x; θ, z) or its conditional distribution to the true conditional p(y|x, θ). Matching the marginal does not constrain the conditional, so the abstract's claim that the method 'ensures correct usage of physics parameters' is not entailed by the training loss. The 'Beyond Marginal Distribution Alignment' paragraph describes an evaluation protocol rather than a training constraint or an identifiability theorem. Table 1b is consistent with residual misspecification even on synthetic data: the OT-GB C2ST values are 0.72±0.06, 0.77±0.02, and 0.57±0.03, far above the reference level of about 0.51. I recommend either adding a conditional consistency term, providing an identifiability analysis, or substantially softening the claim.
  2. [Section 2, Eq. (4)] The displayed weak-OT cost Ck,γ(x, ν) omits the cross term −∫ k(x, y)dν(y) that appears in the unbiased estimator in Eq. (20). Read literally, the cost in Eq. (4) is independent of the coupling between x and y, so Eq. (3) would not implement a conditional OT map at all. This is a central equation and must be corrected so that the theoretical definition matches the algorithm actually used.
  3. [Section 5.3 and Appendix B] The grey-box versus black-box comparison does not isolate the effect of physics knowledge. Grey-box models are NeuralODEs with small convolutional or MLP components, while black-box models are LSTM, 3D-UNet, or ResNet architectures with different capacities and no reported parameter counts. The observed superiority of grey-box models could therefore be due to architecture capacity or optimization rather than the physics inductive bias. Please report capacity-matched or controlled comparisons, for example, the same backbone with and without the fp component.
  4. [Section 2, 'Physics-Guided Optimal Transport Integration'] The claim that fp acts as a 'soft constraint' by narrowing the solution space is plausible but unquantified. Because the training cost does not penalize deviations from the true conditional, the learned component could in principle override fp while still matching the marginal, which is the same failure mode identified in Takeishi & Kalousis (2023) for supervised regression. A small synthetic example showing that different conditionals produce the same marginal, and how the grey-box architecture disambiguates them, would make the argument concrete and would help justify the central claim.
minor comments (4)
  1. [Appendix A.1, Eq. (11)] In the weak-OT formulation, the integral should be with respect to the x-marginal dμ(x); writing dπ(x) is nonstandard and should be clarified.
  2. [Algorithm C, line 10] The gradient update is written as ∂L_g/∂ψ, but the loss was previously denoted L_fψ; the notation should be made consistent.
  3. [Section 3, Chen et al. citation] The sentence about differentiable solvers cites 'Chen et al. (2021)', but the bibliography entry by Chen et al. is the 2018 Neural ODE paper and the 2021 Nature Communications paper is about learning governing equations; the citation appears mismatched.
  4. [Section 4] The phrase 'we provide a more detailed description in the evaluation method described in Section 4' is circular and should be reworded.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the OT training objective and grey-box architecture are independent of the evaluation metrics; self-citations are motivational only.

full rationale

The paper's central construction is not circular. The training objective (Eq. 3 and Algorithm C) is a weak-OT maximin loss that aligns the pushforward marginal with the target marginal using a kernel cost; the evaluation metrics (N-RMSE, C2ST, MMD) are computed on held-out joint samples and do not appear in the training loss, so no fitted input is being renamed as a prediction. The weak-OT theory is imported from Korotin et al. (2023a), an external source with no author overlap, and the grey-box architecture is justified through Neural ODEs and the fixed physics model fp rather than through a self-citation. The self-citations to Takeishi & Kalousis (2021, 2023) are motivational and comparative: they motivate why grey-box architectures can override fp and why regularization may be needed, but the present method does not rely on an unverified uniqueness or existence theorem from those papers for its main derivation. The abstract's claim of 'correct usage of physics parameters' is supported only by evaluation on the joint distribution and by the soft-constraint argument that fp narrows the solution space; this is a conditional-identifiability assumption rather than a circular reduction, since nothing in the training objective forces the joint metric to be minimized. The apparent mismatch between Eq. 4, which omits the cross-term, and the Monte-Carlo estimator Eq. 20, which includes it, is a presentation/consistency concern about whether the stated cost implements the claimed coupling, not a circularity. Overall, the derivation is self-contained against external benchmarks, and the observed limitations are correctness or identifiability risks, not instances of the paper's conclusions being equivalent to its inputs by construction.

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

The central claim rests on the weak-OT framework of Korotin et al. (2023a) and on the grey-box Neural ODE architecture from Takeishi & Kalousis (2021, 2023). The main free choices are the weak-OT regularization parameter γ, the kernel, the latent dimension, and the gradient penalty weight. The paper does not introduce new physical entities. The most consequential axiom is the unproven claim that marginal alignment plus conditioning yields correct joint usage of θ.

free parameters (4)
  • γ (weak-OT variance weight) = 0 for one-to-one, 1 for one-to-many
    Hand-set per task; controls whether the transport map is deterministic or stochastic. The paper uses the endpoints of the allowed interval, which directly affect the optimization and the stochasticity of the map.
  • Choice of kernel for weak-OT cost = distance-based kernel (Frobenius norm)
    Selected heuristically; the theory requires a characteristic positive definite symmetric kernel, but the paper does not verify this property for the distance-based kernel.
  • Latent variable z dimension = same dimension as network input
    Hand-chosen; no sensitivity analysis is provided, and the choice affects the expressiveness of the stochastic map.
  • Gradient penalty coefficient λ = 1
    Fixed to the standard WGAN-GP value; not tuned per task.
assumptions (4)
  • standard math Weak-OT duality and optimality of stochastic maps (Theorem 2 of Korotin et al., 2023a) apply to the conditional/grey-box setting.
    The paper directly transfers the weak-OT Maximin solution to its conditional map Tφ(x; θ, z) without re-deriving the theory for conditioning on θ and z.
  • domain assumption The DGP equals the imperfect physics model fp plus a missing learnable term that the grey-box Neural ODE composition fφ ∘ fp can represent.
    If the model mismatch is not of this functional form, the method cannot complete the physics. The paper constructs experiments so this holds by construction.
  • ad hoc to paper Aligning the marginal ν(y) with a map conditioned on (x, θ) and constrained by fp also aligns the joint distribution π(x, θ, y).
    This is the load-bearing unproven assumption. The training objective (Equation 3, Algorithm C) contains only the marginal target distribution; no joint term is present.
  • domain assumption The optimal transport coupling between source and target marginals corresponds to the true physical coupling.
    OT selects the minimal-distortion coupling, which need not be the physically correct one; the grey-box constraint is expected to mitigate this but is not guaranteed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid Generative Modeling for Incomplete Physics: Deep Grey-Box Meets Optimal Transport." pith.science (2026). https://pith.science/paper/GKRQG5ZQ

@misc{pith2026250622204,
  author       = {Pith},
  title        = {Pith review of: Hybrid Generative Modeling for Incomplete Physics: Deep Grey-Box Meets Optimal Transport},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GKRQG5ZQ}},
  note         = {Machine review of arXiv:2506.22204}
}
read the original abstract

Physics phenomena are often described by ordinary and/or partial differential equations (ODEs/PDEs), and solved analytically or numerically. Unfortunately, many real-world systems are described only approximately with missing or unknown terms in the equations. This makes the distribution of the physics model differ from the true data-generating process (DGP). Using limited and unpaired data between DGP observations and the imperfect model simulations, we investigate this particular setting by completing the known-physics model, combining theory-driven models and data-driven to describe the shifted distribution involved in the DGP. We present a novel hybrid generative model approach combining deep grey-box modelling with Optimal Transport (OT) methods to enhance incomplete physics models. Our method implements OT maps in data space while maintaining minimal source distribution distortion, demonstrating superior performance in resolving the unpaired problem and ensuring correct usage of physics parameters. Unlike black-box alternatives, our approach leverages physics-based inductive biases to accurately learn system dynamics while preserving interpretability through its domain knowledge foundation. Experimental results validate our method's effectiveness in both generation tasks and model transparency, offering detailed insights into learned physics dynamics.

Figures

Figures reproduced from arXiv: 2506.22204 by the authors.

Figure 1
Figure 1. Example of a Reaction Diffusion generation. A visual comparison of the OT black-box and grey-box model generation. The grey-box model shows a generation that closely matches the ground truth. 1 PROBLEM SETTING Consider an unknown data-generating process (DGP) underlying the real world observations, and assume that we have access to some physics-based model fp : X ×Θ 7→ R d , which provides only an imperfect descript… view at source ↗
Figure 2
Figure 2. Probabilistic graphical model of the hybrid model [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Comparison of grey-box and black-box models on different Pendulum settings. Black-box models fail to learn accurate conditional mappings despite ≈ 100M training source samples, while grey-box models excel with the limited simulation budget used in our experiment setting. 6 DISCUSSION Our experimental results demonstrate the importance of physics-informed architecture using grey￾box models, showing particular strengt… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization of trajectory-level and function-level errors of the OT-GB model predictions. The figure compares errors at different scales on the one-to-one Pendulum experiment: trajectory-level deviations (left) and in-solver function evaluation errors (center). The g…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 25 canonical work pages

  1. [1]

    A hybrid simulation of DNN -based gray box models

    Aayushya Agarwal, Yihan Ruan, and Larry Pileggi. A hybrid simulation of DNN -based gray box models. arXiv preprint, 2024

    work page 2024

  2. [2]

    Wasserstein generative adversarial networks

    Martin Arjovsky, Soumith Chintala, and L\' e on Bottou. Wasserstein generative adversarial networks. In Proceedings of the 34th International Conference on Machine Learning - Volume 70, ICML'17, pp.\ 214–223. JMLR.org, 2017

    2017

  3. [3]

    Backhoff-Veraguas, M

    J. Backhoff-Veraguas, M. Beiglb \"o ck, and G. Pammer. Existence, duality, and cyclical monotonicity for weak transport costs. Calculus of Variations and Partial Differential Equations, 58 0 (6): 0 203, November 2019

    work page 2019

  4. [4]

    Applications of weak transport theory

    Julio Backhoff-Veraguas and Gudmund Pammer. Applications of weak transport theory. Bernoulli, 2020

    2020

  5. [5]

    Ricky T. Q. Chen, Yulia Rubanova, Jesse Bettencourt, and David Duvenaud. Neural ordinary differential equations. Advances in Neural Information Processing Systems, 2018

    work page 2018

  6. [6]

    Physics-informed learning of governing equations from scarce data

    Zhao Chen, Yang Liu, and Hao Sun. Physics-informed learning of governing equations from scarce data. Nature Communications, 12 0 (1), October 2021

    work page 2021

  7. [7]

    Kantorovich duality for general transport costs and applications

    Nathael Gozlan, Cyril Roberto, Paul-Marie Samson, and Prasad Tetali. Kantorovich duality for general transport costs and applications. Journal of Functional Analysis, 273 0 (11): 0 3327--3405, 2017

    work page 2017

  8. [8]

    Borgwardt, Malte J

    Arthur Gretton, Karsten M. Borgwardt, Malte J. Rasch, Bernhard Sch \"o lkopf, and Alexander Smola. A kernel two-sample test. Journal of Machine Learning Research, 13 0 (25): 0 723--773, 2012

    work page 2012

Show all 28 references
  1. [9]

    Improved training of wasserstein gans

    Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, and Aaron Courville. Improved training of wasserstein gans. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS'17, pp.\ 5769–5779, Red Hook, NY, USA, 2017. Curran ...

    2017

  2. [10]

    Unifying model-based and neural network feedforward: Physics-guided neural networks with linear autoregressive dynamics

    Johan Kon, Dennis Bruijnen, Jeroen van de Wijdeven, Marcel Heertjes, and Tom Oomen. Unifying model-based and neural network feedforward: Physics-guided neural networks with linear autoregressive dynamics. In 2022 IEEE 61st Conference on Decision and Control (CDC), pp.\ 2475--2...

    2022

  3. [11]

    Kernel neural optimal transport

    Alexander Korotin, Daniil Selikhanovych, and Evgeny Burnaev. Kernel neural optimal transport. In The Eleventh International Conference on Learning Representations, 2023 a

    2023

  4. [12]

    Neural optimal transport

    Alexander Korotin, Daniil Selikhanovych, and Evgeny Burnaev. Neural optimal transport. In The Eleventh International Conference on Learning Representations, 2023 b

    2023

  5. [13]

    Revisiting classifier two-sample tests

    David Lopez-Paz and Maxime Oquab. Revisiting classifier two-sample tests. In International Conference on Learning Representations, 2017

    2017

  6. [14]

    Boyer, Egemen Kolemen, and Jeff G

    Viraj Mehta, Ian Char, Willie Neiswanger, Youngseog Chung, Andrew Oakleigh Nelson, Mark D. Boyer, Egemen Kolemen, and Jeff G. Schneider. Neural dynamical systems: Balancing structure and flexibility in physical prediction. 2021 60th IEEE Conference on Decision and Control (CDC...

    2021

  7. [15]

    Psichogios and Lyle H

    Dimitris C. Psichogios and Lyle H. Ungar. A hybrid neural network-first principles approach to process modeling. AIChE Journal, 38 0 (10): 0 1499--1511, 1992

    1992

  8. [16]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

    Maziar Raissi, Paris Perdikaris, and George E Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378: 0 686--707, 2019

    2019

  9. [17]

    Rico-Martinez, J.S

    R. Rico-Martinez, J.S. Anderson, and I.G. Kevrekidis. Continuous-time nonlinear signal processing: a neural network based approach for gray box identification. In Proceedings of IEEE Workshop on Neural Networks for Signal Processing, pp.\ 596--605, 1994

    1994

  10. [18]

    Generative modeling with optimal transport maps

    Litu Rout, Alexander Korotin, and Evgeny Burnaev. Generative modeling with optimal transport maps. In International Conference on Learning Representations, 2022

    2022

  11. [19]

    MMD aggregated two-sample test

    Antonin Schrab, Ilmun Kim, M \'e lisande Albert, B \'e atrice Laurent, Benjamin Guedj, and Arthur Gretton. MMD aggregated two-sample test. Journal of Machine Learning Research, 24 0 (194): 0 1--81, 2023

    2023

  12. [20]

    Physics-integrated variational autoencoders for robust and interpretable generative modeling

    Naoya Takeishi and Alexandros Kalousis. Physics-integrated variational autoencoders for robust and interpretable generative modeling. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P.S. Liang, and J. Wortman Vaughan (eds.), Advances in Neural Information Processing Systems, volume...

    2021

  13. [21]

    Deep grey-box modeling with adaptive data-driven models toward trustworthy estimation of theory-driven models

    Naoya Takeishi and Alexandros Kalousis. Deep grey-box modeling with adaptive data-driven models toward trustworthy estimation of theory-driven models. In Francisco Ruiz, Jennifer Dy, and Jan-Willem van de Meent (eds.), Proceedings of The 26th International Conference on Artifi...

    2023

  14. [22]

    Thompson and Mark A

    Michael L. Thompson and Mark A. Kramer. Modeling chemical processes using prior knowledge and neural networks. AIChE Journal, 40 0 (8): 0 1328--1340, 1994

    1994

  15. [23]

    p ^3 vae: a physics-integrated generative model

    Romain Thoreau, Laurent Risser, Véronique Achard, Béatrice Berthelot, and Xavier Briottet. p ^3 vae: a physics-integrated generative model. application to the pixel-wise classification of airborne hyperspectral images. arXiv preprint, 2023

    2023

  16. [24]

    Clim ODE : Climate and weather forecasting with physics-informed neural ODE s

    Yogesh Verma, Markus Heinonen, and Vikas Garg. Clim ODE : Climate and weather forecasting with physics-informed neural ODE s. In The Twelfth International Conference on Learning Representations, 2024

    2024

  17. [25]

    C. Villani. Optimal Transport: Old and New. Grundlehren der mathematischen Wissenschaften. Springer Berlin Heidelberg, 2008. ISBN 9783540710509

    2008

  18. [26]

    Robust hybrid learning with expert augmentation

    Antoine Wehenkel, Jens Behrmann, Hsiang Hsu, Guillermo Sapiro, Gilles Louppe, and Joern-Henrik Jacobsen. Robust hybrid learning with expert augmentation. Transactions on Machine Learning Research, 2023

    2023

  19. [27]

    Augmenting physical models with deep networks for complex dynamics forecasting

    Yuan Yin, Vincent Le Guen, Jérémie Dona, Emmanuel de Bézenac, Ibrahim Ayed, Nicolas Thome, and Patrick Gallinari. Augmenting physical models with deep networks for complex dynamics forecasting. Journal of Statistical Mechanics: Theory and Experiment, 2021 0 (12): 0 124012, dec 2021

    2021

  20. [28]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.