Pith. sign in

REVIEW 3 major objections 5 minor 52 references

Physics Informed Constrained Learning of Dynamics from Static Data

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Static snapshots can replace time-course data in physics-informed learning, the paper claims, by deriving fluxes from conservation laws alone when measured derivatives are unavailable.

desk verdict The paper's central claim doesn't survive its own objective: with no data-fitting term, any balanced flux function has zero loss, and the synthetic benchmark is circular; the MPO balancer and real-data pipeline are the genuinely useful pieces. read the letter →

arxiv 2504.12675 v2 pith:TSOQJ5OA submitted 2025-04-17 cs.LG physics.bio-phq-bio.MN

Pengtao Dang , Tingbo Guo , Melissa Fishel , Guang Lin , Wenzhuo Wu , Sha Cao , Chi Zhang This is my paper · ORCID
classification cs.LGphysics.bio-phq-bio.MN
keywords physics-informedneuralnetworksconstrainedlearningmetabolicfluxanalysismessagepassingoptimizationdirectedfactorgraphnon-time-coursedataself-supervisedbalance
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 proposes a learning paradigm called Constrained Learning for estimating the rates or fluxes of a dynamical system from static, non-time-course, and partially observed data. Classical physics-informed neural networks need time-course data and full observation of variables; this work replaces the supervised time-derivative fit with a coherency loss that encodes physical laws, plus a parsimony penalty. For flux-balance systems the physics loss reduces to $\|\Gamma F(D,\Theta)\|^2$, the squared violation of mass conservation at every intermediate node. The authors present MPOCtrL and a message-passing optimizer, and report that on synthetic reaction networks and one real single-cell dataset the method matches or beats existing data-driven flux estimators.

What carries the argument

The central object is a Directed Factor Graph with factor nodes for variables and variable nodes for flux-carrying reactions, connected through a stoichiometric matrix $\Gamma$ that encodes conservation laws. The argument is carried by the combined loss $L = L_\Phi(F(D,\Theta),\Gamma) + L_p(F,\Theta)$: the first term forces predicted fluxes to sit near the kernel of $\Gamma$ (balanced flow), while the second keeps the network simple and interpretable via gating and L2 regularization. A belief-propagation-inspired Message Passing Optimization (MPO) step rebalances variable weights using messages from neighboring nodes, with claimed time and space complexity $O(nK)$.

What would settle it

Take a fixed network and generate synthetic data where the input features are held fixed but the true balanced flux values are randomly permuted across samples, so no feature-flux link exists. If MPOCtrL's predictions do not change under this permutation while the permuted labels differ, the coherency loss alone cannot select the true flux, contradicting the identifiability claim.

Watch

Extended reading notes

Core claim

The central discovery claimed is that the time derivative of a dynamical system can be approximated without any time-course measurements, provided the system's physical laws are encoded as a constraint on a factor-graph representation. Concretely, for mass-carrying flux over a network, the steady-state conservation law gives the coherency loss $L_\Phi = \|\Gamma \{f_1,\ldots,f_K\}^T\|^2$, where $\Gamma$ is the stoichiometric weight matrix. Minimizing this loss together with a parsimony regularizer over feature-driven neural networks yields sample-wise flux estimates. The paper states that on metabolic flux analysis this Constrained Learning approach outperforms all existing data-driven flux estimators.

Load-bearing premise

The load-bearing assumption is that the true flux rates are identifiable from the observed features and the conservation-law constraint alone, even though the loss never compares a predicted flux to any measured flux value.

Editorial extensions

If this is right

  • Physics-informed models could be applied to systems where temporal sampling is impossible, such as human tissue metabolism, using only static snapshots.
  • Cell-wise metabolic flux could be estimated from single-cell RNA-seq data without isotope tracing or time-resolved measurements.
  • The same coherency-loss formulation applies to other mass-carrying networks, including traffic flow, cash flow, and reaction-diffusion systems.
  • The MPO optimizer can serve as a general graph-balancing procedure independent of the neural network, with linear complexity in the number of nodes.
  • The proposed taxonomy of equilibrium versus non-equilibrium systems under steady or dynamic states suggests coherency losses for a broader class of dynamical models.

Reading between the lines

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

  • Because no term in the loss compares predicted flux to measured flux, the method's success on synthetic data likely relies on the constructed nonlinear link between features and balanced flux; when that link is weak, the coherency loss alone cannot distinguish among infinitely many balanced flux functions of the features.
  • A direct identifiability test would be to train on data where true flux is balanced but statistically independent of the input features; predictions should then be arbitrary balanced functions, exposing the hidden dependence on a feature-flux relationship.
  • The real-world validation, with nine metabolites and a single cell line, is narrow relative to the broad superiority claim; broader evaluation with isotope-based flux measurements across tissues would be needed to establish practical advantage.
  • MPO's update rules resemble belief propagation, so its convergence and fixed-point properties on cyclic graphs could be studied formally, potentially extending the framework to loopy networks with guarantees.
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

3 major / 5 minor

Summary. The paper introduces 'Constrained Learning,' a proposed extension of physics-informed neural networks (PINNs) that aims to learn dynamics, specifically flux rates, from static and partially observed data without time-course information. The method represents a dynamic system as a directed factor graph, defines a coherency loss based on mass conservation (||Gamma * F||^2), and combines it with a parsimony regularizer. The authors propose an optimization algorithm, MPOCtrL, that alternates between neural network updates and a message-passing balancing step (MPO). Experiments on synthetic networks and a real scRNA-seq dataset (GSE173433) compare MPOCtrL with scFEA and Compass, and the paper claims that MPOCtrL outperforms existing data-driven flux estimators. The main text and appendix provide algorithmic details, network diagrams, and an appendix section that classifies dynamic systems and proposes coherency losses for various sub-tasks.

Significance. If the central claim were supported, the framework would address a genuine and important limitation of PINNs, namely their reliance on time-course or fully observed data. The paper has some strengths: the implementation is publicly available (GitHub link provided), the appendix gives a systematic taxonomy of dynamic system sub-tasks, and the reported real-world PCC of 0.75 (p=0.019) is a potentially interesting observation. However, the significance is severely undermined by the identifiability issue in the core objective: the loss in Eq. (2) contains no data-fitting term, so any flux function lying in the null space of the stoichiometric matrix achieves zero coherency loss. The synthetic evaluation is also partly circular because the ground-truth fluxes are generated by the same MPO algorithm used during inference. Consequently, the experiments do not establish that the method recovers true flux rates, and the claim of outperforming all existing data-driven flux estimators is not supported.

major comments (3)
  1. [3.4, Eq. (2)] Equation (2) defines the total loss as ||Gamma * F(D, Theta)||^2 + L_p(F, Theta). The observed data D enter only as inputs to the neural networks; there is no term comparing the predicted flux to a measured flux, a measured derivative, or any other observed output. Any function F with Gamma * F(D, Theta) = 0 for all samples achieves zero coherency loss, so the objective cannot distinguish the true flux from any other vector in the null space of the stoichiometric matrix. Since every network with cycles has a nontrivial null space, this is not a corner case. The paper supplies no additional assumption linking the feature distribution to the flux, such as a known kinetic law; Appendix A.6 itself states (Sub-task 3) that for NESS systems the product must be observed to ensure identifiability, and the ESS/unobserved setting used here has no analogous safeguard. Without an identifiability argument or a fitting term, the central claim of Section 3.4 is unsupported.
  2. [A.3.2] The synthetic ground truth is generated by running the MPO algorithm to balance randomly drawn positive numbers on the same directed factor graph that MPOCtrL uses during inference. This makes the benchmark measure self-consistency: the method is only asked to reproduce the fixed point of the same balancing operator used in its own pipeline. A data-driven flux estimator should be evaluated against fluxes that arise from an independent simulation or measurement process. As it stands, the cosine similarities in Table 1 cannot be interpreted as evidence that the method recovers underlying flux rates, and the comparison against scFEA and Compass is correspondingly weakened.
  3. [A.6, Sub-tasks 3 and 4] In the appendix's own taxonomy, identifiability for NESS systems requires that the reaction products (or time-course data) be observed; the loss L_NESS in Eq. (7) explicitly includes a supervised fitting term (X_{k,j} - sum_m F_{m,j})^2. The flux estimation experiments in Section 4 instead assume Vfa = X is completely unobserved (D ∩ Vfa = ∅) and rely solely on Eq. (2), which contains no such fitting term. Thus the paper's experimental setting is precisely the case the appendix indicates is not identifiable, and the claimed approximation of flux from static data is not justified.
minor comments (5)
  1. [Abstract] The word 'scarsity' in the abstract is a typo for 'scarcity'.
  2. [Eq. (2)] The summation index in Eq. (2) runs over j = 1,...,n, but the number of samples is denoted m throughout the paper; this should be corrected.
  3. [Section 3.3 / Algorithm 1] The hyperparameters beta (MPO learning rate) and delta (imbalance loss threshold) are listed in Algorithm 1 but their numerical values or tuning procedure are not specified in the main text or the experiments; this makes reproducibility more difficult.
  4. [Section 3.3] The paragraph beginning 'The message passing step is highly efficient' claims O(nK) time and space complexity, but the appendix (A.1.2) states O(MN). These complexity statements are inconsistent and should be reconciled.
  5. [Figure 1] Figure 1 is referenced in the text but its sub-panels (a), (b), and (c) are not explained in sufficient detail; a reader cannot tell what the geometric illustration or the framework diagram depicts without referring to later sections.

Circularity Check

2 steps flagged · score 6.0 of 10

Eq. 2's coherency-only loss leaves the flux prediction underdetermined, and the synthetic benchmark generates its ground truth with the same MPO operator used at inference.

  1. self definitional [Section 3.2, Definition 2; Section 3.4, Eq. (2)]
    "Constrained Learning identifies functions {F, Θk} = {Fk(Dkj, Θk)} to approximate {fk} by minimizing the following loss function: L = LΦ(F(D, Φ), Γ)+Lp(F, Θ) = Σ_j ||Γ∗{f1(D1j),...,fK(DKj)}^T||^2 + Lp(F, Θ) (2○)"

    The objective contains no term comparing F(D) with any measured flux, derivative, or observed output; the coherency term is zero for every function whose outputs lie in ker(Γ). By construction, any balanced function of D is a minimizer of LΦ, so the loss cannot select the true flux f_k. In the synthetic benchmark the 'true' flux is itself defined as a balanced vector, so the target and the predictor are constrained by the same equation Γf = 0, and no data-dependent term in Eq. (2) connects them.

  2. other [Appendix A.3.2; Algorithm 1 in Appendix A.1.1]
    "To simulate each Yj,., we randomly generate a set of positive numbers and input them into the MPO algorithm to obtain a set of balanced values on a directed factor graph. In other words, for each factor node, its input flux and output flux are equal, and hence balanced."

    The synthetic 'ground truth' flux is produced by running MPO on random positive numbers, while the inference pipeline also uses MPO: Algorithm 1 iterates W ← NNs(GDF,{Xk}) and then W ← (NNs+MPO)(GDF,W,β,N_MPO_max_epoch,α). Thus the evaluation compares two MPO-consistent states on the same directed factor graph. The reported cosine similarities partly measure self-consistency of the MPO balancing operator, not independent recovery of a flux mapping from static features.

full rationale

The most serious issue is in the flux application of Section 3.4. Eq. (2) defines the training loss as the flux-balance residual plus parsimony; there is no term comparing F(D) to a measured flux, a measured derivative, or any other observed output. Every F with ΓF = 0 has the same coherency loss, so the objective is satisfied by any balanced function of the features. The paper supplies no identifiability condition for the case D∩Vfa = ∅; Appendix A.6 explicitly states that for NESS systems 'the product must be observed to ensure identifiability,' but the ESS flux-balance case used here has no analogous safeguard. The synthetic validation is also self-referential: the 'true' flux is generated by running MPO on random numbers, and the predictor itself includes MPO, so high cosine similarities on synthetic data largely measure self-consistency of the MPO balancing operator. The real-world PCC = 0.75 (p = 0.019) against mitochondrial assay data is an external check and is the main independent evidence, but it is a single small nine-metabolite correlation and does not repair the synthetic benchmark circularity or the absence of a data-fitting term in Eq. (2). Because the central flux 'prediction' reduces on the paper's own equations to a projection onto the balance null space, and because the synthetic benchmark is a self-consistency check, the score is 6 rather than lower; the real-world experiment provides some independent content, so the paper is not wholly circular.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

No new physical entities are postulated. The directed factor graph and message variables are computational constructs rather than physical inventions. The load-bearing assumptions are the steady-state balance, feature-to-flux determinism, and unproven convergence of the message-passing balancer.

free parameters (2)
  • MPO learning rate beta = 0.5 used in example; grid {0.1, 0.3, 0.5, 0.7, 0.9} in sensitivity analysis
    Chosen by hand to control the balance between new messages and old variable values in MPO update Eq. (2).
  • Imbalance loss threshold delta = not specified numerically
    Used as a stopping criterion in Algorithms 1 and 2; the value is chosen by hand to avoid the trivial zero solution.
assumptions (3)
  • domain assumption Under steady state or quasi-steady state, the law of conservation of mass gives Gamma * f = 0 for all intermediate metabolites.
    This is the entire coherency loss L_Phi for the flux estimation problem; if the steady-state assumption is violated for any metabolite, the loss is misspecified. Stated in Section 3.4.
  • domain assumption Reaction flux f_k is a deterministic function of the observed features D_kj via a neural network F_k(D_kj, Theta_k).
    The paper assumes gene expression or other static features determine flux, but the loss does not enforce any relationship between features and true flux. Introduced in Definition 2 and Section 3.4.
  • ad hoc to paper MPO converges to the most balanced message flow over the directed factor graph.
    No convergence proof is given; Appendix A.2 argues global perspective and decreasing imbalance loss but does not prove fixed-point convergence. Used in Algorithm 1 and 2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics Informed Constrained Learning of Dynamics from Static Data." pith.science (2026). https://pith.science/paper/TSOQJ5OA

@misc{pith2026250412675,
  author       = {Pith},
  title        = {Pith review of: Physics Informed Constrained Learning of Dynamics from Static Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TSOQJ5OA}},
  note         = {Machine review of arXiv:2504.12675}
}
read the original abstract

A physics-informed neural network (PINN) models the dynamics of a system by integrating the governing physical laws into the architecture of a neural network. By enforcing physical laws as constraints, PINN overcomes challenges with data scarsity and potentially high dimensionality. Existing PINN frameworks rely on fully observed time-course data, the acquisition of which could be prohibitive for many systems. In this study, we developed a new PINN learning paradigm, namely Constrained Learning, that enables the approximation of first-order derivatives or motions using non-time course or partially observed data. Computational principles and a general mathematical formulation of Constrained Learning were developed. We further introduced MPOCtrL (Message Passing Optimization-based Constrained Learning) an optimization approach tailored for the Constrained Learning framework that strives to balance the fitting of physical models and observed data. Its code is available at github link: https://github.com/ptdang1001/MPOCtrL Experiments on synthetic and real-world data demonstrated that MPOCtrL can effectively detect the nonlinear dependency between observed data and the underlying physical properties of the system. In particular, on the task of metabolic flux analysis, MPOCtrL outperforms all existing data-driven flux estimators.

Figures

Figures reproduced from arXiv: 2504.12675 by the authors.

Figure 1
Figure 1. (a) Constrained Learning-based formulation of the flux estimation [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Experiments of MPOCtrL and MPO on Synthetic Data and Real-world [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 46 canonical work pages

  1. [1]

    K. Ahn, J. Zhang, and S. Sra. Understanding the unstable convergence of gradient descent. InInternational Conference on Machine Learning, pages 247–257. PMLR, 2022

    work page 2022

  2. [2]

    Alghamdi, W

    N. Alghamdi, W. Chang, P. Dang, X. Lu, C. Wan, S. Gampala, Z. Huang, J. Wang, Q. Ma, Y. Zang, et al. A graph neural network model to estimate cell-wise metabolic flux using single-cell rna-seq data.Genome research, 31(10):1867–1884, 2021

    work page 2021

  3. [3]

    Arzani, J.-X

    A. Arzani, J.-X. Wang, and R. M. D’Souza. Uncovering near-wall blood flow from sparse data with physics-informed neural networks.Physics of Fluids, 33(7), 2021

    work page 2021

  4. [4]

    Bischof and M

    R. Bischof and M. Kraus. Multi-objective loss balancing for physics-informed deep learning. arXiv preprint arXiv:2110.09813, 2021

    arXiv 2021

  5. [5]

    J. S. Blum, P. A. Wearsch, and P. Cresswell. Pathways of antigen processing. Annual review of immunology, 31:443–473, 2013

    work page 2013

  6. [6]

    S. Cai, Z. Wang, L. Lu, and G. E. Karniadakis. Physics-informed neural networks for inverse problems in nano-optics and metamaterials.Optics Express, 29(8):11815–11828, 2021

    work page 2021

  7. [7]

    Y. Chen, L. Lu, G. E. Karniadakis, and L. Dal Negro. Physics-informed neural networks for inverse problems in nano-optics and metamaterials. Optics express, 28(8):11618–11633, 2020

    work page 2020

  8. [8]

    Cheridito, A

    P. Cheridito, A. Jentzen, A. Riekert, and F. Rossmannek. A proof of convergence for gradient descent in the training of artificial neural networks for constant target functions.Journal of Complexity, 72:101646, 2022

    work page 2022

Show all 52 references
  1. [9]

    I. I. Cplex. V12. 1: User’s manual for cplex.International Business Machines Corporation, 46(53):157, 2009. 16

    2009

  2. [10]

    Daneker, Z

    M. Daneker, Z. Zhang, G. E. Karniadakis, and L. Lu. Systems biology: Identifiability analysis and parameter identification via systems-biology- informed neural networks. InComputational Modeling of Signaling Networks, pages 87–105. Springer, 2023

    2023

  3. [11]

    P. Dang, H. Zhu, T. Guo, C. Wan, T. Zhao, P. Salama, Y. Wang, S. Cao, and C. Zhang. Generalized matrix local low rank representation by random projection and submatrix propagation. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 39...

    2023

  4. [12]

    N. A. Daryakenari, M. De Florio, K. Shukla, and G. E. Karniadakis. Ai- aristotle: A physics-informed framework for systems biology gray-box iden- tification. PLOS Computational Biology, 20(3), 2024

    2024

  5. [13]

    Espinel-Ríos and J

    S. Espinel-Ríos and J. L. Avalos. Hybrid physics-informed metabolic cy- bergenetics: process rates augmented with machine-learning surrogates informed by flux balance analysis. Industrial & Engineering Chemistry Research, 2024

    2024

  6. [14]

    Fang and J

    Z. Fang and J. Zhan. Deep physical informed neural networks for metama- terial design. Ieee Access, 8:24506–24513, 2019

    2019

  7. [15]

    Gharesifard and J

    B. Gharesifard and J. Cortés. Distributed strategies for making a digraph weight-balanced. In2009 47th Annual Allerton Conference on Communica- tion, Control, and Computing (Allerton), pages 771–777. IEEE, 2009

    2009

  8. [16]

    C. S. Henry, M. DeJongh, A. A. Best, P. M. Frybarger, B. Linsay, and R. L. Stevens. High-throughput generation, optimization and analysis of genome-scale metabolic models.Nature biotechnology, 28(9):977–982, 2010

    2010

  9. [17]

    A. T. Ihler, J. W. Fisher III, A. S. Willsky, and D. M. Chickering. Loopy belief propagation: convergence and effects of message errors.Journal of Machine Learning Research, 6(5), 2005

    2005

  10. [18]

    W. Ji, W. Qiu, Z. Shi, S. Pan, and S. Deng. Stiff-pinn: Physics-informed neural network for stiff chemical kinetics.The Journal of Physical Chemistry A, 125(36):8098–8106, 2021

    2021

  11. [19]

    Jiang, D

    X. Jiang, D. Wang, Q. Fan, M. Zhang, C. Lu, and A. P. T. Lau. Physics- informed neural network for nonlinear dynamics in fiber optics.Laser & Photonics Reviews, 16(9):2100483, 2022

    2022

  12. [20]

    G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, and L. Yang. Physics-informed machine learning. Nature Reviews Physics, 3(6):422–440, 2021

    2021

  13. [21]

    K. J. Kauffman, P. Prakash, and J. S. Edwards. Advances in flux balance analysis. Current opinion in biotechnology, 14(5):491–496, 2003. 17

    2003

  14. [22]

    F. R. Kschischang, B. J. Frey, and H.-A. Loeliger. Factor graphs and the sum-product algorithm. IEEE Transactions on information theory, 47(2):498–519, 2001

    2001

  15. [23]

    J. Kuck, S. Chakraborty, H. Tang, R. Luo, J. Song, A. Sabharwal, and S. Er- mon. Belief propagation neural networks.Advances in Neural Information Processing Systems, 33:667–678, 2020

    2020

  16. [24]

    E. M. Lemmon and A. R. Lemmon. High-throughput genomic data in systematics and phylogenetics.Annual Review of Ecology, Evolution, and Systematics, 44:99–121, 2013

    2013

  17. [25]

    Llaneras and J

    F. Llaneras and J. Picó. Stoichiometric modelling of cell metabolism.Journal of bioscience and bioengineering, 105(1):1–11, 2008

    2008

  18. [26]

    L. Lu, X. Meng, Z. Mao, and G. E. Karniadakis. Deepxde: A deep learning library for solving differential equations.SIAM Review, 63(1):208–228, 2021

    2021

  19. [27]

    Y. Ma, D. Tsao, and H.-Y. Shum. On the principles of parsimony and self-consistency for the emergence of intelligence.Frontiers of Information Technology & Electronic Engineering, 23(9):1298–1323, 2022

    2022

  20. [28]

    Mathews, M

    A. Mathews, M. Francisquez, J. W. Hughes, D. R. Hatch, B. Zhu, and B. N. Rogers. Uncovering turbulent plasma dynamics via deep learning from partial observations.Physical Review E, 104(2):025205, 2021

    2021

  21. [29]

    X. Meng, Z. Li, D. Zhang, and G. E. Karniadakis. Composite neural networks that learn from multiscale data: Application to biomedical image segmentation. Journal of Computational Physics, 401:109020, 2020

    2020

  22. [30]

    J. D. Orth, I. Thiele, and B. O. Palsson. Flux balance analysis: cur- rent applications and future directions.Biotechnology and Bioengineering, 107(3):488–496, 2010

    2010

  23. [31]

    Paszke, S

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, et al. Pytorch: An imperative style, high-performance deep learning library.Advances in neural information processing systems, 32, 2019

    2019

  24. [32]

    Priolo, A

    A. Priolo, A. Gasparri, E. Montijano, and C. Sagues. A decentralized algorithm for balancing a strongly connected weighted digraph. In2013 American Control Conference, pages 6547–6552. IEEE, 2013

    2013

  25. [33]

    Pujawan, M

    N. Pujawan, M. M. Arief, B. Tjahjono, and D. Kritchanchai. An integrated shipment planning and storage capacity decision under uncertainty: A simulation study.International Journal of Physical Distribution & Logistics Management, 45(9/10):913–937, 2015. 18

    2015

  26. [34]

    Raissi, P

    M. Raissi, P. Perdikaris, and G. 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, 378:686–707, 2019

    2019

  27. [35]

    Raissi, A

    M. Raissi, A. Yazdani, and G. E. Karniadakis. Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations. Science, 367(6481):1026–1030, 2020

    2020

  28. [36]

    Raman and N

    K. Raman and N. Chandra. Flux balance analysis of biological systems: applications and challenges. Briefings in bioinformatics, 10(4):435–449, 2009

    2009

  29. [37]

    A. I. Rikos, T. Charalambous, and C. N. Hadjicostis. Distributed weight balancing over digraphs.IEEE Transactions on Control of Network Systems, 1(2):190–201, 2014

    2014

  30. [38]

    A. I. Rikos and C. N. Hadjicostis. Distributed balancing of a digraph with integer weights. In52nd IEEE Conference on Decision and Control, pages 1983–1988. IEEE, 2013

    1983

  31. [39]

    Sarabian, H

    M. Sarabian, H. Babaee, and K. Laksari. Physics-informed neural net- works for brain hemodynamic predictions using medical imaging.IEEE transactions on medical imaging, 41(9):2285–2303, 2022

    2022

  32. [40]

    L. Sun, H. Gao, S. Pan, and J.-X. Wang. Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data. Computer Methods in Applied Mechanics and Engineering, 361:112732, 2020

    2020

  33. [41]

    Vardhan and A

    M. Vardhan and A. Randles. Application of physics-based flow models in cardiovascular medicine: Current practices and challenges.Biophysics Reviews, 2(1), 2021

    2021

  34. [42]

    Varma and B

    A. Varma and B. O. Palsson. Stoichiometric flux balance models quanti- tatively predict growth and metabolic by-product secretion in wild-type escherichia coli w3110.Applied and environmental microbiology, 60(10):3724– 3731, 1994

  35. [43]

    Wagner, C

    A. Wagner, C. Wang, J. Fessler, D. DeTomaso, J. Avila-Pacheco, J. Kamin- ski, S. Zaghouani, E. Christian, P. Thakore, B. Schellhaass, et al. Metabolic modeling of single th17 cells reveals regulators of autoimmunity. Cell, 184(16):4168–4185, 2021

    2021

  36. [44]

    Wang and R

    R. Wang and R. Yu. Physics-guided deep learning for dynamical systems: A survey.arXiv preprint arXiv:2107.01272, 2021

    2021 arXiv

  37. [45]

    S. Wang, Y. Teng, and P. Perdikaris. Understanding and mitigating gradient flow pathologies in physics-informed neural networks.SIAM Journal on Scientific Computing, 42(5):A3283–A3310, 2020. 19

    2020

  38. [46]

    S. Wang, X. Yu, and P. Perdikaris. When and why pinns fail to train: A neural tangent kernel perspective.Journal of Computational Physics, 449:110768, 2022

    2022

  39. [47]

    Wiechert, O

    W. Wiechert, O. Schweissgut, H. Takanaga, and W. B. Frommer. Fluxomics: mass spectrometry versus quantitative imaging.Current opinion in plant biology, 10(3):323–330, 2007

    2007

  40. [48]

    Yazdani, L

    A. Yazdani, L. Lu, M. Raissi, and G. E. Karniadakis. Systems biology informed deep learning for inferring parameters and hidden dynamics.PLoS computational biology, 16(11):e1007575, 2020

    2020

  41. [49]

    J. S. Yedidia, W. T. Freeman, Y. Weiss, et al. Understanding belief propagation and its generalizations.Exploring artificial intelligence in the new millennium, 8(236-239):0018–9448, 2003

    2003

  42. [50]

    Zhang, M

    E. Zhang, M. Dao, G. E. Karniadakis, and S. Suresh. Analyses of internal structures and defects in materials using physics-informed neural networks. Science advances, 8(7):eabk0644, 2022

    2022

  43. [51]

    Zhang, Y

    J. Zhang, Y. Zhao, F. Shone, Z. Li, A. F. Frangi, S. Q. Xie, and Z.-Q. Zhang. Physics-informed deep learning for musculoskeletal modeling: Predicting muscle forces and joint kinematics from surface emg.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 31:484–...

    2022

  44. [52]

    Zhang, H

    Z. Zhang, H. Zhu, P. Dang, J. Wang, W. Chang, X. Wang, N. Alghamdi, A. Lu, Y. Zang, W. Wu, et al. Fluxestimator: a webserver for predicting metabolic flux and variations using transcriptomics data.Nucleic Acids Research, 51(W1):W180–W190, 2023. A Appendix / supplemental materi...

    2023

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.