Pith. sign in

REVIEW 4 major objections 3 minor 1 cited by

Physics-Informed Neural Networks with Hard Nonlinear Equality and Inequality Constraints

T0 review · 4 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read KKT-Hardnet makes nonlinear equality and inequality constraints a hard architectural guarantee by solving a KKT projection system inside the network.

desk verdict Sensible KKT projection layer with real experimental gains, but the log-exponential inequality reformulation cannot satisfy complementarity exactly, so the machine-precision claim does not hold. read the letter →

arxiv 2507.08124 v2 pith:TJYJP5A4 submitted 2025-07-10 cs.LG

Ashfaq Iftakher , Rahul Golder , Bimol Nath Roy , M. M. Faruque Hasan This is my paper · ORCID
classification cs.LG MSC 68T0790C3090C33
keywords KKT-Hardnethardconstraintsphysics-informedneuralnetworksdifferentiableprojectionKarush-Kuhn-Tuckerconditionslog-exponentialtransformationFischer-Burmeisterreformulationsurrogatemodeling
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 introduces KKT-Hardnet, a neural network architecture whose predictions always respect known algebraic constraints—equalities and inequalities, linear or nonlinear, and involving both inputs and outputs. After the backbone produces an unconstrained prediction, a differentiable projection layer moves it to the nearest point on the feasible set by solving the Karush–Kuhn–Tucker (KKT) equations of a distance-minimization problem with Newton/Gauss-Newton iterations. Inequality constraints are converted to equalities with slacks, and complementarity is handled by the Fischer–Burmeister equation so that multipliers and slacks stay nonnegative. A log-exponential transformation rewrites many nonlinear constraints as a sparse system of linear and exponential equations, which the Newton solver can handle efficiently. The paper reports that when the projection converges, outputs satisfy the constraints to machine precision or a user-specified tolerance, cutting constraint violations by many orders of magnitude compared with unconstrained MLPs and soft-constrained PINNs.

What carries the argument

The load-bearing object is the differentiable projection layer $\rho(\hat{y})$, defined as the KKT solution of the distance-minimization problem (Eq. 1). It is implemented as a square system $F(y,\lambda)=0$ that combines stationarity, primal feasibility, and complementarity; inequality constraints become equalities through nonnegative slacks, and the complementarity condition is replaced by the Fischer–Burmeister equation, which enforces nonnegativity of slack and dual variables throughout the Newton iterations. The log-exponential transformation restructures a broad class of nonlinear constraints into a sparse form with only linear and exponential terms, so the Jacobian's only nonlinear blocks are diagonal matrices of exponentials. For constraints that are affine in the outputs, the projection collapses to the closed-form orthogonal projection $\hat{y} - B^\top(BB^\top)^{-1}(B\hat{y}+Ax-b)$, avoiding iteration entirely.

What would settle it

Take a trained KKT-Hardnet and evaluate it on inputs where the backbone's raw prediction is a poor initial guess far outside the feasible set; if any output violates a constraint by more than the stated tolerance after the fixed K iterations—say above $10^{-5}$, as already happens in the pooling experiment—then the machine-precision guarantee reduces to a solver-convergence property, not an architectural one.

Watch

Extended reading notes

Core claim

KKT-Hardnet's central claim is that hard constraint satisfaction can be made an architectural property rather than a training objective. Given an input $x$ and a raw prediction $\hat{y}$, the network replaces $\hat{y}$ by the solution of $\min_y \tfrac{1}{2}\|y-\hat{y}\|^2$ subject to $h(x,y)=0$ and $g(x,y)\le 0$, obtained by solving the square KKT system of this projection problem. The solver is unrolled for a fixed number of Newton/Gauss-Newton steps and differentiated through, so gradients flow into the backbone through the projection. Inequalities enter as nonnegative slacks with the complementarity condition $\mu_k s_k=0$ replaced by the Fischer–Burmeister equation $\mu_k+s_k-\sqrt{\mu_k^2+s_k^2}=0$, which automatically keeps multipliers and slacks nonnegative during iteration. The log-exponential transformation converts products, powers, and ratios into exponentials of auxiliary variables, leaving a sparse linear-plus-exponential system whose Jacobian has only diagonal exponential blocks. In the reported experiments, this yields constraint violations around $10^{-6}$–$10^{-9}$, with the nonconvex pooling problem reaching about $10^{-5}$.

Load-bearing premise

The guarantee rests on the assumption that the fixed-budget Newton/Gauss-Newton projection converges to a feasible point of the original nonlinear constraints for every input the network sees, an empirical property rather than a proven theorem.

Editorial extensions

If this is right

  • Networks built this way return outputs that satisfy the constraints at inference time, not just approximately during training.
  • The data-fit and physics-satisfaction objectives are no longer competing, so the penalty-weight tuning that soft-constrained PINNs require disappears.
  • The projection layer is architecture-agnostic, so the same feasible-set projector can wrap any neural backbone, including convolutional or recurrent networks.
  • When constraints are affine in the outputs, the projection is a single analytic linear layer, making feasibility essentially free.
  • Feasible component surrogates should not propagate constraint violations when cascaded into larger process models.

Reading between the lines

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

  • If the fixed Newton budget is what actually limits hardness, the 'machine precision' claim should be read as 'solver-convergence precision'; the architecture guarantees feasibility only to the extent the projection solver converges within K steps.
  • The log-exponential transformation's nonnegativity restriction suggests a natural extension for signed outputs via difference-of-two-exponentials splitting, at the cost of extra variables and conditioning.
  • Because the projector is differentiable and architecture-agnostic, the same layer could serve as a structural regularizer in low-data regimes for image or sequence models, a use the paper does not develop.
  • The implicit-differentiation adjoint sketched in Appendix B points toward memory-efficient training of much deeper projection layers, a practical route the paper leaves partly open.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The paper introduces KKT-Hardnet, a neural network architecture that augments a standard MLP backbone with a differentiable projection layer. The projection solves the KKT conditions of a distance-minimization problem to enforce nonlinear equality and inequality constraints on the network outputs, and the authors additionally propose a log-exponential transformation that rewrites general nonlinear constraints into a structured system of linear and exponential equations. The method is demonstrated on three illustrative examples, a pooling problem, and an extractive-distillation simulation, with comparisons against unconstrained MLPs and soft-constrained PINNs. The paper claims that the architecture enforces constraints 'up to machine precision' during both training and inference.

Significance. If the central claims were valid, the paper would make a useful contribution to hard-constrained scientific machine learning: the KKT projection idea is sensible, the analytic projection for affine-in-output constraints is clean and well presented, the code is made publicly available, and the comparison against MLP/PINN baselines with a penalty-weight sweep in Appendix A is thorough. The method also targets practically important nonlinear constraints in chemical-process systems. However, the key advertised guarantee—machine-precision enforcement of inequality constraints via the log-exponential transformation—is mathematically inconsistent, and the experimental results contradict the abstract's blanket claim. These are load-bearing issues, not presentation problems.

major comments (4)
  1. [Section 2.2 and Example 3, Eqs. (52d)-(52h)] The log-exponential reformulation of inequality constraints is internally inconsistent. In Example 3, equations F5 and F8 enforce mu_I = exp(z3) and s = exp(z5), so both mu_I and s are strictly positive for any finite z3, z5. But the Fischer-Burmeister complementarity condition, written as F4 (z8 - mu_I - s = 0) with z8 = sqrt(mu_I^2 + s^2), requires sqrt(mu_I^2 + s^2) = mu_I + s, which has no solution when both arguments are strictly positive. Consequently the transformed KKT system has no exact real solution; the reported residual of 1e-9 in Table 3 is a least-squares artifact, not the result of an exact projection. This directly invalidates the claim that the log-exponential transformation enforces inequality constraints to machine precision.
  2. [Section 3.3, data generation and Table 3] The regression problem in Example 3 is infeasible as stated: the data are generated as y = x^2 with x sampled uniformly from [1,2], while the enforced inequality is y - x <= 0. For every x > 1, the target x^2 lies strictly outside the feasible set. The projection therefore maps the network output to the boundary y = x, and the KKT-Hardnet MSE in Table 3 is 1.059, roughly 35 times larger than the unconstrained MLP MSE of 3.038e-2. The text claims KKT-Hardnet achieves 'comparable MSE to the unconstrained MLP,' which is contradicted by the table. This example cannot support the paper's claims about the benefits of hard inequality enforcement.
  3. [Abstract and Table 6] The abstract states that KKT-Hardnet 'enforces linear and nonlinear equality and inequality constraints up to machine precision,' but the reported experiments do not support this. The pooling problem in Table 6 reports a mean absolute constraint violation of about 1.05e-5, and Example 3 in Table 3 reports 1e-9; both are many orders of magnitude above double-precision machine epsilon (~2e-16). Even the analytic affine projection in Table 5 reports 8.66e-8. The conclusion's softer wording, 'within specified tolerance and machine precision,' should replace the abstract's unconditional claim, and the paper should state explicitly which experiments meet which tolerance.
  4. [Section 2.1 and Algorithm 1] The 'hard' guarantee is not supported by the numerical scheme. The projection layer runs a fixed number K of Newton/Gauss-Newton iterations (K = 30 or 100) with no convergence theorem, no failure analysis, and no bound on the feasibility residual after K steps. The text itself acknowledges in Section 3.4.2 that the final residual may lie anywhere between the stopping tolerance and the numerical precision limit. As a result, feasibility is an empirical property of the iterates on the tested problems, not a guarantee by construction. The paper should either provide a convergence or residual-bound theorem or substantially qualify the meaning of 'hard' throughout the title, abstract, and introduction.
minor comments (3)
  1. [Section 3.2, Eq. (36c)] Equation (36c) reads '3x7 = x8 = 0,' which is inconsistent with the intended relation 3x7 - x8 = 0 used in the augmented system; this appears to be a typographical error.
  2. [Remark 1 and Abstract] Remark 1 restricts the log-exponential transformation to nonnegative outputs y in R^p_{\ge 0}, but this limitation is not mentioned in the abstract or the contributions list, where the method is presented as general. This structural restriction should be stated prominently.
  3. [Section 3.4.2, text near Figure 7] The statement that 'no violation spikes above 10^-6' is consistent with the figure, but the subsequent sentence attributes oscillations to 'solver termination under finite precision'; the observed residual floor near 1e-6 is the stopping tolerance, not machine precision, and the wording should be adjusted to avoid conflating the two.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: KKT-Hardnet's feasibility guarantee is an architectural construction, and the reported residuals are empirical solver outputs, not fitted predictions.

full rationale

KKT-Hardnet's feasibility guarantee is not a fitted prediction: the projected output is defined as the solution of the KKT system (Eq. 11) associated with the distance-minimization problem (Eq. 1), so if the Newton/Gauss-Newton iterations converge, the equality and inequality constraints hold by construction rather than by inference from data. This is an architectural guarantee, not a claim that the constraints are derived from the network. The constraints themselves are supplied as domain knowledge (mass balances, reflux relation, sulfur specifications) and are not tuned to the targets ybar; the numerical violations reported (3.5e-8 to 1e-5) are empirical residuals of finite-iteration solves, so they carry independent content. Self-citations ([2], [9], [10], [22], [44]) support process-simulation details and prior surrogate-modeling work, but none is load-bearing for the KKT projection derivation or for the benchmark comparisons; external works such as ENFORCE [39], OptNet [37], DC3 [35], and Chen et al. [26] provide the surrounding context. The log-exponential transformation is a variable substitution; presenting it as a way to maintain nonnegativity is a modeling choice, not a circular reuse of the conclusion. One correctness caveat, distinct from circularity: in Example 3, Eqs. 52e and 52h force muI = exp(z3) > 0 and s = exp(z5) > 0, which makes the Fischer-Burmeister equation F4 = z8 - muI - s = 0 with z8 = sqrt(muI^2 + s^2) unsatisfiable exactly; the reported 1e-9 violation is therefore an approximate least-squares residual rather than machine precision. This undermines the headline claim but does not make the derivation circular.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central claim rests on the solver converging and on the log-exp transformation being valid for the problem class. The free parameters are all solver hyperparameters, not physical constants fitted to data. The auxiliary z variables are computational bookkeeping. The main limiting assumption is the nonnegativity requirement and the reliance on local Newton convergence.

free parameters (4)
  • K (number of Newton steps) = K=30 for most examples, K=100 for pooling
    Number of Gauss-Newton iterations in the projection layer; set by hand, not tuned automatically. Larger K improves feasibility but increases memory cost due to unrolled backpropagation.
  • gamma (Tikhonov regularization) = 1e-3 (process case), 1e-2 (pooling case)
    Regularization parameter added to the Gauss-Newton normal equations to ensure invertibility of the Jacobian; chosen by hand per problem.
  • alpha (step length) = 0.5 fixed in pooling, Armijo backtracking in process case
    Step length in the Newton update; set manually or via line search, influencing convergence speed and stability.
  • epsilon (slack/multiplier initialization offset) = 1e-3
    Initialization offset used in Example 3 to keep slack and multiplier variables positive at the start of Newton iterations.
assumptions (4)
  • domain assumption The constraint functions h and g are differentiable, and the number of equality constraints is at most the number of outputs (NE <= p).
    Stated in Section 2, Assumptions 1 and 2, to ensure the KKT system is square and solvable.
  • domain assumption The constraints are feasible and linearly independent.
    Section 2, Assumption 2; ensures the projection problem has a solution and the Jacobian has full rank (at least locally).
  • ad hoc to paper For the log-exponential transformation, all input and output variables are nonnegative (x in R^m, y in R^p_>=0).
    Remark 1 restricts the transformation to nonnegative variables; this excludes problems with signed outputs, a limitation not emphasized in the abstract.
  • standard math Newton/Gauss-Newton iteration converges locally to a solution of the KKT system when the Jacobian is invertible.
    Used in Section 2.1 to justify the projection layer; standard local convergence theory, but no global convergence guarantees are provided for nonconvex constraints.
invented entities (1)
  • Auxiliary log-exp variables z (e.g., z1,...,z10)
    purpose: Rewrite nonlinear terms (powers, products, ratios) as exponentials to obtain a sparse linear plus exponential system (Eq. 12).
    They are purely computational auxiliaries introduced in Section 2.2; they have no physical meaning or external falsifiable handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics-Informed Neural Networks with Hard Nonlinear Equality and Inequality Constraints." pith.science (2026). https://pith.science/paper/TJYJP5A4

@misc{pith2026250708124,
  author       = {Pith},
  title        = {Pith review of: Physics-Informed Neural Networks with Hard Nonlinear Equality and Inequality Constraints},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TJYJP5A4}},
  note         = {Machine review of arXiv:2507.08124}
}
read the original abstract

Traditional physics-informed neural networks (PINNs) do not guarantee strict constraint satisfaction. This is problematic in engineering systems where minor violations of governing laws can degrade the reliability and consistency of model predictions. In this work, we introduce KKT-Hardnet, a neural network architecture that enforces linear and nonlinear equality and inequality constraints up to machine precision. It leverages a differentiable projection onto the feasible region by solving Karush-Kuhn-Tucker (KKT) conditions of a distance minimization problem. Furthermore, we reformulate the nonlinear KKT conditions via a log-exponential transformation to construct a sparse system with linear and exponential terms. We apply KKT-Hardnet to nonconvex pooling problem and a real-world chemical process simulation. Compared to multilayer perceptrons and PINNs, KKT-Hardnet achieves strict constraint satisfaction. It also circumvents the need to balance data and physics residuals in PINN training. This enables the integration of domain knowledge into machine learning towards reliable hybrid modeling of complex systems.

Figures

Figures reproduced from arXiv: 2507.08124 by the authors.

Figure 1
Figure 1. Neural network architecture of KKT-Hardnet for hard constrained machine learning. Unconstrained outputs [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Projection of neural net outputs onto the constraint manifold. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Learning curve for example 1. Left: Normalised RMSE (RMSE is normalized by the range of outputs), Right: [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Learning curves for example 2. Left: Normalized RMSE; Right: absolute constraint violation over 1200 [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Learning curves for example 3. Left: Normalized RMSE; Right: absolute constraint violation (log scale). [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]
Figure 6
Figure 6. Figure 6: Extractive distillation of R-410A using an Ionic liquid [EMIM][SCN]. The input variables are denoted in red. [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]
Figure 7
Figure 7. Figure 7: Learning curves for the simulation of extractive distillation-based simulation of R-410A using an ionic liquid [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: Learning curves for the analytic projection case. Left: RMSE; Right: absolute constraint violation (log scale). [PITH_FULL_IMAGE:figures/full_fig_p019_8.png]
Figure 9
Figure 9. Figure 9: Process flowsheet of a pooling problem. The product can be formed from three feeds. The input variables are [PITH_FULL_IMAGE:figures/full_fig_p020_9.png]
Figure 10
Figure 10. Figure 10: Learning curves for the pooling problem case study. Left: RMSE; Right: absolute constraint violation over [PITH_FULL_IMAGE:figures/full_fig_p021_10.png]
Figure 11
Figure 11. Figure 11: Training curves comparing MLP, KKT-Hardnet, and PINNs with a sweep over [PITH_FULL_IMAGE:figures/full_fig_p025_11.png]
Figure 12
Figure 12. Figure 12: Training curves comparing MLP, KKT-Hardnet, and PINNs with a sweep over [PITH_FULL_IMAGE:figures/full_fig_p026_12.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Physics-Constrained Machine Learning for Chemical Engineering

    cs.LG 2025-08 unverdicted novelty 2.0 of 10

    A perspective on physics-constrained machine learning for chemical engineering that summarizes approaches, applications, and open challenges without introducing a new method.

Reference graph

Works this paper leans on

45 extracted references · 27 canonical work pages · cited by 1 Pith paper

  1. [1]

    Physics- informed machine learning

    George Em Karniadakis, Ioannis G Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang. Physics- informed machine learning. Nature Reviews Physics, 3(6):422–440, 2021

    2021

  2. [2]

    M Faruque Hasan

    Ashfaq Iftakher, Ty Leonard, and M. M Faruque Hasan. Integrating different fidelity models for process optimization: A case of equilibrium and rate-based extractive distillation using ionic liquids. Computers & Chemical Engineering, 192:108890, 2025

    work page 2025

  3. [3]

    Physics-informed neural networks for solving reynolds-averaged navier–stokes equations

    Hamidreza Eivazi, Mojtaba Tahani, Philipp Schlatter, and Ricardo Vinuesa. Physics-informed neural networks for solving reynolds-averaged navier–stokes equations. Physics of Fluids, 34(7), 2022

    work page 2022

  4. [4]

    Theory-guided hard constraint projection (hcp): A knowledge-based data-driven scientific machine learning method

    Yuntian Chen, Dou Huang, Dongxiao Zhang, Junsheng Zeng, Nanzhe Wang, Haoran Zhang, and Jinyue Yan. Theory-guided hard constraint projection (hcp): A knowledge-based data-driven scientific machine learning method. Journal of Computational Physics, 445:110624, 2021

    work page 2021

  5. [5]

    Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations

    Maziar Raissi, Alireza Yazdani, and George Em Karniadakis. Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations. Science, 367(6481):1026–1030, 2020

    work page 2020

  6. [6]

    Real-time optimal control via deep neural networks: study on landing problems

    Carlos Sánchez-Sánchez and Dario Izzo. Real-time optimal control via deep neural networks: study on landing problems. Journal of Guidance, Control, and Dynamics , 41(5):1122–1135, 2018

    work page 2018

  7. [7]

    Industrial, large-scale model predictive control with structured neural networks

    Pratyush Kumar, James B Rawlings, and Stephen J Wright. Industrial, large-scale model predictive control with structured neural networks. Computers & chemical engineering, 150:107291, 2021

    work page 2021

  8. [8]

    Advances in surrogate based modeling, feasibility analysis, and optimization: A review

    Atharv Bhosekar and Marianthi Ierapetritou. Advances in surrogate based modeling, feasibility analysis, and optimization: A review. Computers & Chemical Engineering , 108:250–267, 2018

    work page 2018

Show all 45 references
  1. [9]

    Ashfaq Iftakher, Chinmay M Aras, Mohammed Sadaf Monjur, and M. M. Faruque Hasan. Data-driven approxima- tion of thermodynamic phase equilibria. AIChE Journal, 68(6):e17624, 2022

    2022

  2. [10]

    Ashfaq Iftakher, Chinmay M Aras, Mohammed Sadaf Monjur, and M. M. Faruque Hasan. Guaranteed error- bounded surrogate modeling and application to thermodynamics. In Computer Aided Chemical Engineering , volume 49, pages 1831–1836. Elsevier, 2022

    2022

  3. [11]

    Overview of surrogate modeling in chemical process engineering

    Kevin McBride and Kai Sundmacher. Overview of surrogate modeling in chemical process engineering. Chemie Ingenieur Technik, 91(3):228–239, 2019

    2019

  4. [12]

    Formulating data-driven surrogate models for process optimization

    Ruth Misener and Lorenz Biegler. Formulating data-driven surrogate models for process optimization. Computers & Chemical Engineering, 179:108411, 2023

    2023

  5. [13]

    The alamo approach to machine learning

    Zachary T Wilson and Nikolaos V Sahinidis. The alamo approach to machine learning. Computers & Chemical Engineering, 106:785–795, 2017

    2017

  6. [14]

    Learning surrogate models for simulation-based optimization

    Alison Cozad, Nikolaos V Sahinidis, and David C Miller. Learning surrogate models for simulation-based optimization. AIChE Journal, 60(6):2211–2227, 2014

    2014

  7. [15]

    Multilayer feedforward networks are universal approxi- mators

    Kurt Hornik, Maxwell Stinchcombe, and Halbert White. Multilayer feedforward networks are universal approxi- mators. Neural networks, 2(5):359–366, 1989

    1989

  8. [16]

    Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. Communications of the ACM, 60(6):84–90, 2017

    2017

  9. [17]

    Generative adversarial networks

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Communications of the ACM, 63(11):139–144, 2020

    2020

  10. [18]

    Deep learning

    Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. Deep learning. nature, 521(7553):436–444, 2015

    2015

  11. [19]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems , 30, 2017. 22

    2017

  12. [20]

    A tutorial review of neural network modeling approaches for model predictive control

    Yi Ming Ren, Mohammed S Alhajeri, Junwei Luo, Scarlett Chen, Fahim Abdullah, Zhe Wu, and Panagiotis D Christofides. A tutorial review of neural network modeling approaches for model predictive control. Computers & Chemical Engineering, 165:107956, 2022

    2022

  13. [21]

    Physics-informed neural networks (pinns) for fluid mechanics: A review

    Shengze Cai, Zhiping Mao, Zhicheng Wang, Minglang Yin, and George Em Karniadakis. Physics-informed neural networks (pinns) for fluid mechanics: A review. Acta Mechanica Sinica, 37(12):1727–1738, 2021

    2021

  14. [22]

    Multiscale high-throughput screening of ionic liquid solvents for mixed-refrigerant separation

    Ashfaq Iftakher, Mohammed Sadaf Monjur, Ty Leonard, Rafiqul Gani, and MM Faruque Hasan. Multiscale high-throughput screening of ionic liquid solvents for mixed-refrigerant separation. Computers & Chemical Engineering, 199:109138, 2025

    2025

  15. [23]

    Physics-informed machine learning: case studies for weather and climate modelling

    Karthik Kashinath, M Mustafa, Adrian Albert, JL Wu, C Jiang, Soheil Esmaeilzadeh, Kamyar Azizzadenesheli, R Wang, Ashesh Chattopadhyay, A Singh, et al. Physics-informed machine learning: case studies for weather and climate modelling. Philosophical Transactions of the Royal So...

    2021

  16. [24]

    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, 378:686–707, 2019

    2019

  17. [25]

    Physics-informed neural networks for heat transfer problems

    Shengze Cai, Zhicheng Wang, Sifan Wang, Paris Perdikaris, and George Em Karniadakis. Physics-informed neural networks for heat transfer problems. Journal of Heat Transfer, 143(6):060801, 2021

    2021

  18. [26]

    Physics-informed neural networks with hard linear equality constraints

    Hao Chen, Gonzalo E Constante Flores, and Can Li. Physics-informed neural networks with hard linear equality constraints. Computers & Chemical Engineering , 189:108764, 2024

    2024

  19. [27]

    When and why pinns fail to train: A neural tangent kernel perspective

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

    2022

  20. [28]

    Data-driven strategies for optimization of integrated chemical plants

    Kaiwen Ma, Nikolaos V Sahinidis, Satyajith Amaran, Rahul Bindlish, Scott J Bury, Devin Griffith, and Sreekanth Rajagopalan. Data-driven strategies for optimization of integrated chemical plants. Computers & Chemical Engineering, 166:107961, 2022

    2022

  21. [29]

    Characterizing possible failure modes in physics-informed neural networks

    Aditi Krishnapriyan, Amir Gholami, Shandian Zhe, Robert Kirby, and Michael W Mahoney. Characterizing possible failure modes in physics-informed neural networks. Advances in neural information processing systems , 34:26548–26560, 2021

    2021

  22. [30]

    Is L2 physics-informed loss always suitable for training physics-informed neural networks? Advances in Neural Information Processing Systems , 35:8278–8290, 2022

    Chuwei Wang, Shanda Li, Di He, and Liwei Wang. Is L2 physics-informed loss always suitable for training physics-informed neural networks? Advances in Neural Information Processing Systems , 35:8278–8290, 2022

    2022

  23. [31]

    Challenges in training pinns: A loss landscape perspective

    Pratik Rathore, Weimu Lei, Zachary Frangella, Lu Lu, and Madeleine Udell. Challenges in training pinns: A loss landscape perspective. arXiv preprint arXiv:2402.01868, 2024

    2024 arXiv

  24. [32]

    Hard-constrained neural networks with universal approxi- mation guarantees

    Youngjae Min, Anoopkumar Sonar, and Navid Azizan. Hard-constrained neural networks with universal approxi- mation guarantees. arXiv preprint arXiv:2410.10807, 2024

    2024

  25. [33]

    Imposing hard constraints on deep networks: Promises and limitations

    Pablo Márquez-Neila, Mathieu Salzmann, and Pascal Fua. Imposing hard constraints on deep networks: Promises and limitations. arXiv preprint arXiv:1706.02025, 2017

    2017 arXiv

  26. [34]

    Enforcing analytic constraints in neural networks emulating physical systems

    Tom Beucler, Michael Pritchard, Stephan Rasp, Jordan Ott, Pierre Baldi, and Pierre Gentine. Enforcing analytic constraints in neural networks emulating physical systems. Physical review letters, 126(9):098302, 2021

    2021

  27. [35]

    Dc3: A learning method for optimization with hard constraints

    Priya L Donti, David Rolnick, and J Zico Kolter. Dc3: A learning method for optimization with hard constraints. arXiv preprint arXiv:2104.12225, 2021

    2021 arXiv

  28. [36]

    Numerical optimization

    Jorge Nocedal and Stephen J Wright. Numerical optimization. Springer, 1999

    1999

  29. [37]

    Optnet: Differentiable optimization as a layer in neural networks

    Brandon Amos and J Zico Kolter. Optnet: Differentiable optimization as a layer in neural networks. In International conference on machine learning , pages 136–145. PMLR, 2017

    2017

  30. [38]

    Differentiable convex optimization layers

    Akshay Agrawal, Brandon Amos, Shane Barratt, Stephen Boyd, Steven Diamond, and J Zico Kolter. Differentiable convex optimization layers. Advances in neural information processing systems , 32, 2019

    2019

  31. [39]

    Enforce: Exact nonlinear constrained learning with adaptive-depth neural projection

    Giacomo Lastrucci and Artur M Schweidtmann. Enforce: Exact nonlinear constrained learning with adaptive-depth neural projection. arXiv preprint arXiv:2502.06774, 2025

    2025 arXiv

  32. [40]

    Smoothed fischer-burmeister equation methods for the complementarity problem

    Houyuan Jiang. Smoothed fischer-burmeister equation methods for the complementarity problem. Department of Mathematics, The University of Melbourne (Australia , 1997

    1997

  33. [41]

    Pytorch: An imperative style, high-performance deep learning library

    A Paszke. Pytorch: An imperative style, high-performance deep learning library. arXiv preprint arXiv:1912.01703, 2019

    1912 arXiv

  34. [42]

    Adam: A method for stochastic optimization

    Diederik P Kingma. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014. 23

    2014 arXiv

  35. [43]

    The moore–penrose pseudoinverse: A tutorial review of the theory

    João Carlos Alves Barata and Mahir Saleh Hussein. The moore–penrose pseudoinverse: A tutorial review of the theory. Brazilian Journal of Physics, 42:146–165, 2012

    2012

  36. [44]

    Mohammed Sadaf Monjur, Ashfaq Iftakher, and M. M. Faruque Hasan. Separation process synthesis for high-gwp refrigerant mixtures: Extractive distillation using ionic liquids. Industrial & Engineering Chemistry Research , 61(12):4390–4406, 2022

    2022

  37. [45]

    Handbook of test problems in local and global optimization, volume 33

    Christodoulos A Floudas, Panos M Pardalos, Claire Adjiman, William R Esposito, Zeynep H Gümüs, Stephen T Harding, John L Klepeis, Clifford A Meyer, and Carl A Schweiger. Handbook of test problems in local and global optimization, volume 33. Springer Science & Business Media, 2...

    2013

Pith tools

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