Pith. sign in

REVIEW 3 major objections 7 minor 47 references

Reviewed by Pith at T0; open to challenge.

T0 means a machine referee read the full paper against a public rubric. The mark states how deep the mechanical check went, never who wrote it. the ladder, T0–T4 →

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

T0 review · glm-5.2

PDE solution families compress to 2-4 latent dimensions

2026-07-08 08:45 UTC pith:W6MLCT27

load-bearing objection Clean invariance proofs and a novel Fourier-shell decomposition of latent covariance, but the headline claim of training-independent observables is not empirically validated — no multi-run comparison exists. the 3 major comments →

arxiv 2607.06348 v1 pith:W6MLCT27 submitted 2026-07-07 cs.LG cs.NAmath.NAphysics.comp-ph

Physics-Informed Neural Embeddings of PDE Solution Families

Raul Jimenez , Svitlana Mayboroda , Pavlos Protopapas , Leonid Sarieddine , David N. Spergel , Pedro Taranc\'on-\'Alvarez This is my paper
classification cs.LG cs.NAmath.NAphysics.comp-ph
keywords physics-informed neural networkspartial differential equationslatent spaceprincipal component analysisBurgers equationreduced-order modelingsolution manifoldsspectral decomposition
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper introduces a multihead physics-informed neural network architecture in which a shared body learns a latent manifold representing the solution space of a PDE, while linear heads reconstruct individual solutions for different initial conditions. The key innovation is a head-orthogonalization penalty that removes degeneracies in the latent representation, making the principal-component spectrum of the latent space stable across independent training runs. When applied to the 1D viscous Burgers equation (with heat and wave equations as controls), the method reveals that only 2-4 principal components capture approximately 95% of the latent-space variance despite a nominal embedding dimension of 20. The authors further define Fourier-shell spectral weights that decompose each principal component's variance across wavenumber bands, yielding a scale-resolved profile of the learned manifold that is invariant under the change-of-basis freedom left by the orthogonalization penalty. Together, these two observables -- the PCA spectrum and the Fourier-shell spectral weights -- are proposed as training-independent diagnostics of solution-manifold geometry that emerge directly from PDE constraints, without requiring precomputed solution datasets.

Core claim

The central discovery is that the latent space learned by a physics-informed neural network under PDE constraints exhibits pronounced low-dimensional organization: for Burgers dynamics, only 2-4 principal components capture roughly 95% of the variance in a 20-dimensional embedding. This compression is not imposed architecturally but emerges from the simultaneous enforcement of the governing equation across a family of initial conditions. Crucially, because the initial condition is built into the network ansatz by construction, the principal components measure the residual variability the network must learn to evolve solutions away from their initial profiles -- not the full solution variance

What carries the argument

The multihead PINN with linear heads and a head-orthogonalization penalty (Eq. 10) is the central architectural object. The orthogonalization drives the head-weight matrix W to be approximately orthogonal, which makes the change-of-basis matrix A = W^{-1} what-hat-W between two training runs approximately orthogonal, stabilizing the PCA eigenvalue spectrum. The Fourier-shell spectral weight rho_n(Q) = v_n^T C^(Q) v_n is the central invariant observable, decomposing each principal component's eigenvalue into scale-resolved contributions that survive latent reparametrizations.

Load-bearing premise

The claim that the PCA spectrum and spectral weights are training-independent rests on the head-orthogonalization penalty driving the head-weight matrix W to be approximately orthogonal, so that the change-of-basis matrix A between two training runs is close to orthogonal. The paper provides a theoretical error bound depending on the orthogonalization error delta = ||A^T A - I||_2, but never reports numerical values of delta for any experiment, leaving the practical validity,

What would settle it

Train the same multihead PINN on Burgers equation with multiple random seeds, measure the orthogonalization error delta = ||A^T A - I||_2 for each pair of runs, and check whether the PCA eigenvalue differences |lambda_k(C) - lambda_k(C-hat)| are bounded by the Appendix D formula. If delta is not small or the eigenvalue differences exceed the bound, the training-independence claim is unsupported.

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

If this is right

  • If the PCA spectrum and spectral weights are genuinely training-independent, they could serve as reproducible diagnostic tools for comparing solution-manifold geometry across different PDEs, initial-condition families, and viscosity regimes without requiring shared training infrastructure.
  • The low-dimensional compression suggests that equation-informed reduced-order models could be constructed directly from the leading principal components, bypassing the snapshot-based approach of classical POD or DMD methods.
  • Extending the framework to Navier-Stokes or higher-dimensional nonlinear PDEs would test whether the observed hierarchical compression is a generic property of nonlinear dynamics or specific to 1D Burgers, heat, and wave equations.
  • The Fourier-shell decomposition could provide a data-driven analog of scale-dependent coarse-graining, potentially connecting learned latent hierarchies to renormalization-group structure in turbulent systems.

Where Pith is reading between the lines

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

  • The error bound in Appendix D depends on the orthogonalization error delta = ||A^T A - I||_2 being small, but no numerical values of delta are reported for any experiment. If delta is not small in practice, the claimed invariance of the PCA spectrum and spectral weights across training runs rests on an unverified premise. A direct test would report delta values alongside the PCA spectra for each e
  • The bound also involves unquantified constants (K_H1, ||H||_{L2}, ||W^{-1}||_2), making it impossible to assess whether the theoretical guarantee is tight enough to matter for the reported loss values (10^{-4} to 10^{-2}). Without these constants, the bound serves as a qualitative argument rather than a quantitative guarantee.
  • The dependence of effective dimensionality and spectral profiles on the initial-condition family suggests the learned manifold reflects properties of the sampled solution family rather than an intrinsic property of the PDE itself. This raises the question of whether a universal manifold exists for a given PDE or whether the geometry is fundamentally conditional on the IC ensemble.
  • The 10-day training cost per experiment on an H100 GPU for 1D problems raises practical concerns about scalability to 2D or 3D systems, where the collocation-point count and network capacity would need to grow substantially.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. This paper introduces a multihead physics-informed neural network (PINN) framework for learning finite-dimensional embeddings of PDE solution families. A shared body network learns a latent representation, while linear heads reconstruct individual solutions for different initial conditions. A head-orthogonalization penalty is introduced to remove degeneracies in the latent space, and the authors prove that this stabilizes the PCA spectrum across training realizations. The paper also introduces a Fourier-shell decomposition that assigns scale-resolved spectral weights to each principal component, with proven invariance under latent reparametrizations. The method is applied to the 1D viscous Burgers equation, with the heat and wave equations as robustness checks. The main empirical finding is that only 2–4 principal components capture approximately 95% of the latent-space variance despite a nominal dimension of 20.

Significance. The paper makes a genuine methodological contribution by introducing training-invariant observables (PCA spectrum and Fourier-shell spectral weights) for PINN latent spaces. The covariance transformation law (Eq. 8), the invariance of eigenvalues under orthogonal transformations, and the Fourier-shell decomposition (Sec. 3.4) are mathematically sound. The error bound in Appendix D, while not tight, correctly identifies the key quantities controlling invariance. The framework is falsifiable: the claims about low-dimensional structure and spectral profiles are concrete and testable. The code is stated to be publicly available. The application to three distinct PDE types (dissipative nonlinear, dissipative linear, non-dissipative linear) with three IC families provides a reasonable breadth of evidence for the dimensional reduction phenomenon.

major comments (3)
  1. §3.3 and §3.4, abstract, and §1: The central claim that the PCA spectrum and spectral weights ρ_n(Q) are 'reproducible across independent training runs' and 'training-independent observables' is supported only by the theoretical bound in Appendix D. I can find no figure or table showing PCA spectra or ρ_n(Q) profiles overlaid for two or more independently trained models. Every result in §4 and Figs. 3–4 appears to come from a single training run per equation/IC-family combination. The theoretical bound involves δ = ||A^T A - I||_2, the loss residuals ε, ε̂, and constants K_H1, ||H||_L2, ||W^{-1}||_2, none of which are numerically evaluated. Without at least two independent runs showing agreement, or numerical values of δ demonstrating the bound is small in practice, the headline claim of training independence rests entirely on an unverified theoretical guarantee. This is the load-bearing
  2. issue for the paper's central contribution. At minimum, the authors should report δ values for their experiments and ideally overlay PCA spectra from 2–3 independent runs for at least one equation/IC-family combination.
  3. Table 1, wave equation rows: The minimum loss for the wave equation with polynomial ICs is 4.72×10^{-2} and with wavelet ICs is 7.62×10^{-2}. These are relatively large residuals (roughly 5–8% of typical solution amplitudes of order 1). The paper does not discuss whether the PCA spectra and spectral weights reported for these cases are reliable at this level of loss residual, given that the error bound in Appendix D scales with √ε + √ε̂. The reader cannot assess whether the dimensional reduction claims for the wave equation are robust or artifacts of insufficient convergence. The authors should either demonstrate that these loss levels are sufficient for the bound to be meaningful, or flag these results as preliminary.
minor comments (7)
  1. §3.2.2, Eq. (10): The penalty weight λ = 1×10^{-3} is mentioned in Appendix A but not in the main text where Eq. (10) is introduced. Consider cross-referencing.
  2. Fig. 3 caption: The shell boundaries are listed as 'Q1 := 0.00' which is unclear; presumably Q1 corresponds to the zero-wavenumber (DC) mode. This should be stated explicitly.
  3. §3.4: The invariance proof for ρ_n(Q) assumes non-degenerate eigenvalues (stated near the end of the section). The paper should comment on whether this assumption holds in practice for the reported experiments, particularly when eigenvalues are close (as may occur for higher-index PCs).
  4. Appendix D: The H^1_x stability estimate involving K_H1 is stated without proof or reference. A citation or brief justification would strengthen the rigor of the bound.
  5. Table 1: The 'Dom. shells (PC1-PC3)' column lists Q3,Q2,Q4 for Burgers/Fourier but the formatting is ambiguous — it is unclear whether these correspond to PC1, PC2, PC3 respectively. A clearer format (e.g., PC1:Q3, PC2:Q2, PC3:Q4) would help.
  6. §4: The text states 'absolute errors of order 10^{-2}' but Table 1 reports minimum losses ranging from 10^{-4} to 10^{-2}. Clarify whether the figure refers to pointwise absolute error (from Figs. 2, C1, C2) or the loss function values.
  7. The AI disclaimer section is commendable for transparency. No change needed, but the reference to Villaescusa-Navarro et al. (2025) (the Denario project) should be verified for correctness as it appears to be a very recent preprint.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for a careful and constructive report. The referee correctly identifies that our central claim of training-independence lacks empirical verification through multi-run comparisons and numerical evaluation of the error bound quantities. We agree this must be addressed. On the wave equation loss levels, we acknowledge the concern and will add discussion of reliability. We outline concrete revisions below.

read point-by-point responses

  1. Referee: The central claim that the PCA spectrum and spectral weights ρ_n(Q) are 'reproducible across independent training runs' and 'training-independent observables' is supported only by the theoretical bound in Appendix D. No figure or table shows PCA spectra or ρ_n(Q) profiles overlaid for two or more independently trained models. Every result in §4 and Figs. 3–4 appears to come from a single training run per equation/IC-family combination. The theoretical bound involves δ, ε, ε̂, and constants none of which are numerically evaluated. Without at least two independent runs showing agreement, or numerical values of δ demonstrating the bound is small in practice, the headline claim of training independence rests entirely on an unverified theoretical guarantee.

    Authors: The referee is correct on all counts. Our manuscript currently presents the theoretical invariance argument (Sec. 3.3–3.4) and the error bound (Appendix D) but does not provide the empirical counterpart: overlaid PCA spectra and ρ_n(Q) profiles from independently trained models, nor numerical values of δ, ε, and ε̂. This is a genuine gap between our claims and our evidence, and we will address it in the revision. Specifically, we commit to the following changes: (1) We will train 2–3 independent models (different random seeds) for at least the Burgers/Fourier and Burgers/wavelet combinations, and overlay the resulting PCA spectra and ρ_n(Q) profiles in a new figure. These experiments are currently underway; each requires approximately 10 days on an H100. (2) We will report numerical values of δ = ||A^T A - I||_2, ε, and ε̂ for all equation/IC-family combinations in a new table, so the reader can assess whether the bound in Appendix D is small in practice. (3) We will soften the language in the abstract and §1 from 'reproducible across independent training runs' to a more precise statement: the observables are provably invariant under orthogonal reparametrizations (exact result), and the error bound (Appendix D) quantifies the deviation when the orthogonality and loss residuals are finite; empirical verification across independent runs is provided in the revised results. We agree that without the multi-run overlay and the δ values, the training-independence claim is not adequately supported, and we will not make the claim in its current unqualified form. revision: yes

  2. Referee: Table 1, wave equation rows: The minimum loss for the wave equation with polynomial ICs is 4.72×10^{-2} and with wavelet ICs is 7.62×10^{-2}. These are relatively large residuals (roughly 5–8% of typical solution amplitudes of order 1). The paper does not discuss whether the PCA spectra and spectral weights reported for these cases are reliable at this level of loss residual, given that the error bound in Appendix D scales with √ε + √ε̂. The authors should either demonstrate that these loss levels are sufficient for the bound to be meaningful, or flag these results as preliminary.

    Authors: The referee raises a valid concern. The wave equation losses are indeed higher than those for Burgers and heat, and we did not discuss the implications for the reliability of the PCA spectra in those cases. We will address this in two ways. First, once we report numerical values of δ, ε, and ε̂ as committed in our response to the first comment, the reader will be able to directly assess whether the bound is meaningful for the wave equation cases. Second, we will add an explicit discussion in §4 noting that the wave equation results—particularly for polynomial and wavelet ICs—have higher loss residuals and that the dimensional reduction claims for those cases should be interpreted with appropriate caution. If the multi-run overlay (also committed above) shows that the PCA spectra are stable across seeds despite the higher loss, this will provide direct empirical evidence of reliability; if not, we will flag those specific results as preliminary, as the referee suggests. We agree that the current manuscript is silent on this issue and that it should not be. revision: yes

Circularity Check

0 steps flagged

No significant circularity. The invariance proof is a genuine mathematical derivation from the orthogonalization penalty design; the dimensional reduction is an empirical finding not forced by construction.

full rationale

The paper's central claims are not circular. The invariance of the PCA spectrum and Fourier-shell spectral weights under latent reparametrizations is proven mathematically (§3.3, Eq. 8; §3.4, Eq. 11): if A = W^{-1}Ŵ is orthogonal, then C = AĈA^T is an orthogonal similarity transform, so eigenvalues are preserved. The head-orthogonalization penalty (Eq. 10) is designed to make W approximately orthogonal, which makes A approximately orthogonal — but this is a design choice that enables the invariance, not a circular argument where the conclusion is assumed. The proof that ρ_n(Q) is invariant (§3.4) follows from the transformation properties of C^(Q) and is a genuine derivation. The low-dimensional structure (2–4 PCs capturing 95% variance) is an empirical finding from training, not forced by any definition or fit. The error bound in Appendix D is a legitimate perturbation analysis involving δ, loss residuals, and regularity constants. The self-citations (Tarancón-Álvarez et al. 2025 for MH architecture; Villaescusa-Navarro et al. 2025 for the AI suggestion) are not load-bearing for the mathematical claims — the invariance proofs are self-contained in this paper. The absence of multi-run empirical validation is a correctness/verification gap, not a circularity issue: the theoretical argument is derived from first principles, not assumed. Score 2 reflects one minor self-citation for the MH architecture that is not load-bearing for the paper's central mathematical results.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 0 invented entities

No new physical entities, particles, forces, or dimensions are introduced. The method operates within standard PINN and PCA frameworks.

free parameters (4)
  • n_b (latent dimension) = 20
    Chosen by the authors; not derived from the PDE or data. The effective dimensionality result depends on this choice.
  • λ (orthogonalization penalty weight) = 1e-3
    Stated as found experimentally in Appendix A; no systematic hyperparameter search reported.
  • Network architecture (5 layers, 128 neurons) = 5×128
    Chosen by the authors; affects the capacity and thus the learned latent space.
  • Number of training epochs = 3.0e5
    Fixed across experiments; convergence at this epoch count is assumed but not verified.
axioms (4)
  • domain assumption The head weight matrix W is full rank (invertible in the square case)
    Invoked in Sec. 3.2.2 to derive the latent transformation law (Eq. 7). The paper states this is 'expected when the heads correspond to sufficiently diverse ICs/BCs' but does not verify it numerically.
  • domain assumption The orthogonalization penalty makes W approximately orthogonal (δ small)
    The entire framework for training-independent observables depends on A = W^{-1}Ŵ being approximately orthogonal. No numerical value of δ is reported.
  • domain assumption The H^1_x stability estimate holds with constant K_H1
    Used in Appendix D to bound the eigenvalue difference. The constant depends on regularity of ψ and ψ̂ but is never quantified.
  • domain assumption Eigenvalues of the latent covariance are non-degenerate
    Assumed in Sec. 3.4 to ensure eigenvectors transform as v_n = A v̂_n. Degenerate eigenvalues would break the uniqueness of eigenvector assignment.

pith-pipeline@v1.1.0-glm · 26247 in / 3984 out tokens · 299024 ms · 2026-07-08T08:45:18.567835+00:00 · methodology

0 comments
read the original abstract

We introduce a physics-informed framework for learning finite-dimensional embeddings of solution families of partial differential equations. The method uses a multihead Physics-Informed Neural Network in which a shared body learns a latent manifold representing the solution space, while linear heads reconstruct individual solutions associated with different initial conditions. A head-orthogonalization penalty removes degeneracies in the latent representation and stabilizes the principal-component spectrum across training realizations. Because the initial condition is built into the network output by construction, these principal components measure the additional variability the network learns on top of the initial profile, not the full solution itself. We apply the method to the one-dimensional viscous Burgers equation, with the heat and wave equations as robustness checks. For a latent dimension $n_b=20$, the learned manifolds exhibit pronounced effective dimensional reduction: for Burgers dynamics, only $2$-$4$ principal components capture about $95\%$ of the latent-space variance, while $4$-$7$ capture about $99\%$, depending on the initial-condition family; the same qualitative compression holds for the heat and wave equations. We also split the wavenumber axis into bands (``Fourier shells'') and measure how much each band contributes to every principal component. The resulting frequency profile is invariant under the change-of-basis freedom that the orthogonalization penalty leaves in the latent space, and is therefore reproducible across independent training runs. More broadly, this establishes the learned spectral profiles and principal components as robust observables of solution-manifold geometry.

discussion (0)

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

Reference graph

Works this paper leans on

47 extracted references · 47 canonical work pages · 5 internal anchors

  1. [1]

    Physics Reports 767--769:1--101

    Alexakis A, Biferale L (2018) Cascades and transitions in turbulent flows. Physics Reports 767--769:1--101. doi:10.1016/j.physrep.2018.08.001

    work page doi:10.1016/j.physrep.2018.08.001 2018

  2. [2]

    P., Hill, D

    Angulo RE, White SDM (2010) One simulation to fit them all --- changing the background parameters of a cosmological n-body simulation. Mon Not Roy Astron Soc 405:143--154. doi:10.1111/j.1365-2966.2010.16459.x

    work page doi:10.1111/j.1365-2966.2010.16459.x 2010

  3. [3]

    arXiv preprint arXiv:251111137

    Auroy S, Protopapas P (2025) One-shot transfer learning for nonlinear pdes with perturbative pinns. arXiv preprint arXiv:251111137

    work page 2025

  4. [4]

    Gravitational Duals from Equations of State

    Bea Y, Jimenez R, Mateos D, et al (2024) Gravitational duals from equations of state . JHEP 07:087. doi:10.1007/JHEP07(2024)087, https://arxiv.org/abs/2403.14763 arXiv:2403.14763 [hep-th]

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1007/jhep07(2024)087 2024

  5. [5]

    IEEE Trans Pattern Anal Mach Intell 35:1798--1828

    Bengio Y, Courville A, Vincent P (2013) Representation learning: A review and new perspectives. IEEE Trans Pattern Anal Mach Intell 35:1798--1828

    work page 2013

  6. [6]

    Annual Review of Fluid Mechanics 25:539--575

    Berkooz G, Holmes P, Lumley JL (1993) The proper orthogonal decomposition in the analysis of turbulent flows. Annual Review of Fluid Mechanics 25:539--575. doi:10.1146/annurev.fl.25.010193.002543

    work page doi:10.1146/annurev.fl.25.010193.002543 1993

  7. [7]

    Annals of Mathematics 189(1):101--144

    Buckmaster T, Vicol V (2019) Nonuniqueness of weak solutions to the N avier- S tokes equation. Annals of Mathematics 189(1):101--144. doi:10.4007/annals.2019.189.1.3

    work page doi:10.4007/annals.2019.189.1.3 2019

  8. [8]

    Adv Appl Mech 1:171--199

    Burgers JM (1948) A mathematical model illustrating the theory of turbulence. Adv Appl Mech 1:171--199

    work page 1948

  9. [9]

    Journal of Computational Physics 395:166--185

    Cai W, Jiang C, Wang Y, et al (2019) Structure-preserving algorithms for the two-dimensional sine-gordon equation with neumann boundary conditions. Journal of Computational Physics 395:166--185. doi:10.1016/j.jcp.2019.05.048

    work page doi:10.1016/j.jcp.2019.05.048 2019

  10. [10]

    Journal of Open Source Software 5(46):1931

    Chen F, Sondak D, Protopapas P, et al (2020) Neurodiffeq: A python package for solving differential equations with neural networks. Journal of Open Source Software 5(46):1931

    work page 2020

  11. [11]

    Physical Review E 54(1):376

    Chen LY, Goldenfeld N, Oono Y (1996) Renormalization group and singular perturbations: Multiple scales, boundary layers, and reductive perturbation theory. Physical Review E 54(1):376

    work page 1996

  12. [12]

    J Sci Comput 92:88

    Cuomo S, Schiano Di Cola V, Giampaolo F, et al (2022) Scientific machine learning through physics-informed neural networks: Where we are and what's next. J Sci Comput 92:88

    work page 2022

  13. [13]

    One-Shot Transfer Learning of Physics-Informed Neural Networks

    Desai S, Mattheakis M, Joy H, et al (2022) One-shot transfer learning of physics-informed neural networks. arXiv preprint arXiv:211011286 doi:10.48550/arXiv.2110.11286, iCML AI4Science Workshop 2022

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2110.11286 2022

  14. [14]

    Communications in Numerical Methods in Engineering 10(3):195--201

    Dissanayake MWMG, Phan‐Thien N (1994) Neural‐network‐based approximations for solving partial differential equations. Communications in Numerical Methods in Engineering 10(3):195--201. doi:10.1002/cnm.1640100303

    work page doi:10.1002/cnm.1640100303 1994

  15. [15]

    Cambridge University Press

    Doering CR, Gibbon JD (1995) Applied Analysis of the Navier--Stokes Equations. Cambridge University Press

    work page 1995

  16. [16]

    Enflo BO, Hedberg CM (2002) Theory of Nonlinear Acoustics in Fluids. No. 67 in Fluid Mechanics and Its Applications, Kluwer Academic Publishers, Dordrecht

    work page 2002

  17. [17]

    Clay Mathematics Institute Millennium Prize Problem description, ://www.claymath.org/millennium-problems/navier-stokes-equation

    Fefferman CL (2000) Existence and smoothness of the N avier-- S tokes equation. Clay Mathematics Institute Millennium Prize Problem description, ://www.claymath.org/millennium-problems/navier-stokes-equation

    work page 2000

  18. [18]

    Computer Methods in Applied Mechanics and Engineering 424:116906

    Ferrer-Sánchez A, Martín-Guerrero JD, de Austri-Bazan RR, et al (2024) Gradient-annihilated pinns for solving riemann problems: Application to relativistic hydrodynamics. Computer Methods in Applied Mechanics and Engineering 424:116906. doi:10.1016/j.cma.2024.116906

    work page doi:10.1016/j.cma.2024.116906 2024

  19. [19]

    Solving Differential Equations Using Neural Network Solution Bundles

    Flamant C, Protopapas P, Sondak D (2020) Solving differential equations using neural network solution bundles. CoRR abs/2006.14372. ://arxiv.org/abs/2006.14372, https://arxiv.org/abs/2006.14372 2006.14372

    work page internal anchor Pith review Pith/arXiv arXiv 2020

  20. [20]

    Phys Rev A 16:732--749

    Forster D, Nelson DR, Stephen MJ (1977) Large-distance and long-time properties of a randomly stirred fluid. Phys Rev A 16:732--749

    work page 1977

  21. [21]

    Frisch U (1995) Turbulence: The Legacy of A. N. Kolmogorov. Cambridge University Press

    work page 1995

  22. [22]

    CRC Press

    Goldenfeld N (2018) Lectures on Phase Transitions and the Renormalization Group. CRC Press

    work page 2018

  23. [23]

    MIT Press

    Goodfellow I, Bengio Y, Courville A (2016) Deep Learning. MIT Press

    work page 2016

  24. [24]

    arXiv:220705748

    Goswami S, Bora A, Yu Y, et al (2022) Physics-informed deep neural operator networks. arXiv:220705748

    work page 2022

  25. [25]

    Commun Pure Appl Math 3:201--230

    Hopf E (1950) The partial differential equation u_t + uu_x = u_ xx . Commun Pure Appl Math 3:201--230

    work page 1950

  26. [26]

    Nat Rev Phys 3:422--440

    Karniadakis GE, Kevrekidis IG, Lu L, et al (2021) Physics-informed machine learning. Nat Rev Phys 3:422--440

    work page 2021

  27. [27]

    IEEE Transactions on Neural Networks 11(5):1041--1049

    Lagaris I, Likas A, Papageorgiou D (2000) Neural-network methods for boundary value problems with irregular boundaries. IEEE Transactions on Neural Networks 11(5):1041--1049. doi:10.1109/72.870037

    work page doi:10.1109/72.870037 2000

  28. [28]

    Course of theoretical physics / by L

    Landau LD, Lifshitz EM (1987) Fluid Mechanics, Second Edition: Volume 6 (Course of Theoretical Physics), 2nd edn., Butterworth-Heinemann. Course of theoretical physics / by L. D. Landau and E. M. Lifshitz, Vol. 6, ://www.worldcat.org/isbn/0750627670

    work page arXiv 1987

  29. [29]

    In: Analysis and numerics of partial differential equations

    Lassila T, Manzoni A, Quarteroni A, et al (2013) Generalized reduced basis methods and n-width estimates for the approximation of the solution manifold of parametric pdes. In: Analysis and numerics of partial differential equations. Springer, p 307--329

    work page 2013

  30. [30]

    arXiv:201008895

    Li Z, Kovachki N, Azizzadenesheli K, et al (2020) Fourier neural operator for parametric partial differential equations. arXiv:201008895

    work page 2020

  31. [31]

    J Mach Learn Res 24:1--26

    Li Z, Huang DZ, Liu B, et al (2023) Fourier neural operator with learned deformations for pdes on general geometries. J Mach Learn Res 24:1--26

    work page 2023

  32. [32]

    arXiv preprint arXiv:250212177

    Liu S, Protopapas P, Sondak D, et al (2025) Recent advances of neurodiffeq--an open-source library for physics-informed neural networks. arXiv preprint arXiv:250212177

    work page 2025

  33. [33]

    Nat Mach Intell 3:218--229

    Lu L, Jin P, Pang G, et al (2021 a ) Learning nonlinear operators via deeponet based on the universal approximation theorem of operators. Nat Mach Intell 3:218--229

    work page 2021

  34. [34]

    SIAM review 63(1):208--228

    Lu L, Meng X, Mao Z, et al (2021 b ) Deepxde: A deep learning library for solving differential equations. SIAM review 63(1):208--228

    work page 2021

  35. [35]

    Physical Symmetries Embedded in Neural Networks

    Mattheakis M, Protopapas P, Sondak D, et al (2019) Physical symmetries embedded in neural networks. ://arxiv.org/abs/1904.08991, https://arxiv.org/abs/1904.08991 arXiv:1904.08991

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  36. [36]

    Hamiltonian neural networks for solving equations of motion

    Mattheakis M, Sondak D, Dogra AS, et al (2020) Hamiltonian neural networks for solving equations of motion. arXivorg perpetual, non-exclusive license doi:10.48550/ARXIV.2001.11107

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2001.11107 2020

  37. [37]

    Nonlinear Dynamics 41(1--3):309--325

    Mezi \'c I (2005) Spectral properties of dynamical systems, model reduction and decompositions. Nonlinear Dynamics 41(1--3):309--325. doi:10.1007/s11071-005-2824-x

    work page doi:10.1007/s11071-005-2824-x 2005

  38. [38]

    arXiv preprint arXiv:221100214

    Pellegrin R, Bullwinkel B, Mattheakis M, et al (2022) Transfer learning with physics-informed neural networks for efficient simulation of branched flows. arXiv preprint arXiv:221100214

    work page 2022

  39. [39]

    Springer

    Quarteroni A, Manzoni A, Negri F (2015) Reduced basis methods for partial differential equations: an introduction. Springer

    work page 2015

  40. [40]

    Raissi and P

    Raissi M, Perdikaris P, Karniadakis G (2019) Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics 378:686--707. doi:10.1016/j.jcp.2018.10.045

    work page doi:10.1016/j.jcp.2018.10.045 2019

  41. [41]

    Dynamic mode decomposition of numerical and experimental data

    Schmid PJ (2010) Dynamic mode decomposition of numerical and experimental data. Journal of Fluid Mechanics 656:5--28. doi:10.1017/S0022112010001217

    work page doi:10.1017/s0022112010001217 2010

  42. [42]

    Journal of Computational Physics 375:1339--1364

    Sirignano J, Spiliopoulos K (2018) Dgm: A deep learning algorithm for solving partial differential equations. Journal of Computational Physics 375:1339--1364. doi:10.1016/j.jcp.2018.08.029

    work page doi:10.1016/j.jcp.2018.08.029 2018

  43. [43]

    Communications Physics 8(1):335

    Taranc \'o n- \'A lvarez P, Tejerina-P \'e rez P, Jimenez R, et al (2025) Efficient pinns via multi-head unimodular regularization of the solutions space. Communications Physics 8(1):335

    work page 2025

  44. [44]

    ://arxiv.org/abs/2510.26887, https://arxiv.org/abs/2510.26887 arXiv:2510.26887

    Villaescusa-Navarro F, Bolliet B, Villanueva-Domingo P, et al (2025) The denario project: Deep knowledge ai agents for scientific discovery. ://arxiv.org/abs/2510.26887, https://arxiv.org/abs/2510.26887 arXiv:2510.26887

    work page arXiv 2025

  45. [45]

    Astrophys J 915:71

    Villaescusa-Navarro F, et al (2021) The camels project: Cosmology and astrophysics with machine-learning simulations. Astrophys J 915:71

    work page 2021

  46. [46]

    Physics reports 12(2):75--199

    Wilson KG, Kogut J (1974) The renormalization group and the expansion. Physics reports 12(2):75--199

    work page 1974

  47. [47]

    Journal of Computational Physics 425:109913

    Yang L, Meng X, Karniadakis GE (2021) B-pinns: Bayesian physics-informed neural networks for forward and inverse pde problems with noisy data. Journal of Computational Physics 425:109913

    work page 2021