Investigating and Mitigating Failure Modes in Physics-Informed Neural Networks (PINNs)

Investigating and Mitigating Failure Modes in Physics-Informed Neural Networks (PINNs)

Year:    2023

Author:    Shamsulhaq Basir

Communications in Computational Physics, Vol. 33 (2023), Iss. 5 : pp. 1240–1269

Abstract

This paper explores the difficulties in solving partial differential equations (PDEs) using physics-informed neural networks (PINNs). PINNs use physics as a regularization term in the objective function. However, a drawback of this approach is the requirement for manual hyperparameter tuning, making it impractical in the absence of validation data or prior knowledge of the solution. Our investigations of the loss landscapes and backpropagated gradients in the presence of physics reveal that existing methods produce non-convex loss landscapes that are hard to navigate. Our findings demonstrate that high-order PDEs contaminate backpropagated gradients and hinder convergence. To address these challenges, we introduce a novel method that bypasses the calculation of high-order derivative operators and mitigates the contamination of backpropagated gradients. Consequently, we reduce the dimension of the search space and make learning PDEs with non-smooth solutions feasible. Our method also provides a mechanism to focus on complex regions of the domain. Besides, we present a dual unconstrained formulation based on Lagrange multiplier method to enforce equality constraints on the model’s prediction, with adaptive and independent learning rates inspired by adaptive subgradient methods. We apply our approach to solve various linear and non-linear PDEs.

You do not have full access to this article.

Already a Subscriber? Sign in as an individual or via your institution

Journal Article Details

Publisher Name:    Global Science Press

Language:    English

DOI:    https://doi.org/10.4208/cicp.OA-2022-0239

Communications in Computational Physics, Vol. 33 (2023), Iss. 5 : pp. 1240–1269

Published online:    2023-01

AMS Subject Headings:    Global Science Press

Copyright:    COPYRIGHT: © Global Science Press

Pages:    30

Keywords:    Constrained optimization Lagrangian multiplier method Stokes equation convection equation convection-dominated convection-diffusion equation heat transfer in composite medium Lid-driven cavity problem.

Author Details

Shamsulhaq Basir

  1. A principled distance-aware uncertainty quantification approach for enhancing the reliability of physics-informed neural network

    Li, Jinwu | Long, Xiangyun | Deng, Xinyang | Jiang, Wen | Zhou, Kai | Jiang, Chao | Zhang, Xiaoge

    Reliability Engineering & System Safety, Vol. 245 (2024), Iss. P.109963

    https://doi.org/10.1016/j.ress.2024.109963 [Citations: 2]
  2. k-space physics-informed neural network (k-PINN) for compressed spectral mapping and efficient inversion of vibrations in thin composite laminates

    Hedayatrasa, Saeid | Fink, Olga | Van Paepegem, Wim | Kersemans, Mathias

    Mechanical Systems and Signal Processing, Vol. 223 (2025), Iss. P.111920

    https://doi.org/10.1016/j.ymssp.2024.111920 [Citations: 0]