Physics-informed neural networks for approximating dynamic (hyperbolic) PDEs of second order in time: error analysis and algorithms

DOI10.1016/J.JCP.2023.112527MaRDI QIDQ6087958zbMATH OpenFDO

Authors Yanxia Qian, Yong-Chao Zhang, Yunqing Huang, S. Dong

Publication date 16 November 2023

Published in Journal of Computational Physics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2303.12245

zbMATH Keywords

error estimateneural networkPDEPINNscientific machine learning

Mathematics Subject Classification ID

Stability and convergence of numerical methods for initial value and initial-boundary value problems involving PDEs (65M12)

Abstract: We consider the approximation of a class of dynamic partial differential equations (PDE) of second order in time by the physics-informed neural network (PINN) approach, and provide an error analysis of PINN for the wave equation, the Sine-Gordon equation and the linear elastodynamic equation. Our analyses show that, with feed-forward neural networks having two hidden layers and the $anh$ activation function, the PINN approximation errors for the solution field, its time derivative and its gradient field can be effectively bounded by the training loss and the number of training data points (quadrature points). Our analyses further suggest new forms for the training loss function, which contain certain residuals that are crucial to the error estimate but would be absent from the canonical PINN loss formulation. Adopting these new forms for the loss function leads to a variant PINN algorithm. We present ample numerical experiments with the new PINN algorithm for the wave equation, the Sine-Gordon equation and the linear elastodynamic equation, which show that the method can capture the solution well.

Recommendations

• Error analysis for physics-informed neural networks (PINNs) approximating Kolmogorov PDEs
• A Rate of Convergence of Physics Informed Neural Networks for the Linear Second Order Elliptic PDEs
• Error estimates for physics-informed neural networks approximating the Navier-Stokes equations
• On the convergence of physics informed neural networks for linear second-order elliptic and parabolic type PDEs
• Numerical analysis of physics-informed neural networks and related models in physics-informed machine learning

Cites work

• A method for representing periodic functions and enforcing exactly periodic boundary conditions with deep neural networks
• A modified batch intrinsic plasticity method for pre-training the random coefficients of extreme learning machines
• A unified scalable framework for causal sweeping strategies for physics-informed neural networks (PINNs) and their temporal decompositions
• Analysis of the generalization error: empirical risk minimization over deep artificial neural networks overcomes the curse of dimensionality in the numerical approximation of Black-Scholes partial differential equations
• Classical elastodynamics as a linear symmetric hyperbolic system
• Classical Global Solutions for Non-linear Klein-Gordon-Schrödinger Equations
• Conservative physics-informed neural networks on discrete domains for conservation laws: applications to forward and inverse problems
• Deep learning architectures. A mathematical approach
• Deep least-squares methods: an unsupervised learning-based numerical method for solving elliptic PDEs
• Deep Neural Network Approximation Theory
• DeepXDE: a deep learning library for solving differential equations
• DGM: a deep learning algorithm for solving partial differential equations
• Enhancing Accuracy of Deep Learning Algorithms by Training with Low-Discrepancy Sequences
• Error analysis for physics-informed neural networks (PINNs) approximating Kolmogorov PDEs
• Error estimates for deep learning methods in fluid dynamics
• Estimates on the generalization error of physics-informed neural networks for approximating a class of inverse problems for PDEs
• Estimates on the generalization error of physics-informed neural networks for approximating PDEs
• Extended Physics-Informed Neural Networks (XPINNs): A Generalized Space-Time Domain Decomposition Based Deep Learning Framework for Nonlinear Partial Differential Equations
• Extreme learning machine collocation for the numerical solution of elliptic PDEs with sharp gradients
• Generalization bounds for function approximation from scattered noisy data
• Global existence of classical solutions to systems of nonlinear wave equations with different speeds of propagation
• Global existence of small solutions to nonlinear evolution equations
• Greedy training algorithms for neural networks and applications to PDEs
• High-order discontinuous Galerkin methods for the elastodynamics equation on polygonal and polyhedral meshes
• Higher-order error estimates for physics-informed neural networks approximating the primitive equations
• scientific article; zbMATH DE number 5297089 (Why is no real title available?)
• scientific article; zbMATH DE number 5060482 (Why is no real title available?)
• scientific article; zbMATH DE number 3244852 (Why is no real title available?)
• Infinite-dimensional dynamical systems in mechanics and physics.
• Local extreme learning machines and domain decomposition for solving linear and nonlinear partial differential equations
• Machine learning approximation algorithms for high-dimensional fully nonlinear partial differential equations and second-order backward stochastic differential equations
• Meta-learning PINN loss functions
• MgNet: a unified framework of multigrid and convolutional neural network
• Normal forms and quadratic nonlinear Klein-Gordon equations
• Numerical approximation of partial differential equations by a variable projection method with artificial neural networks
• Numerical solution and bifurcation analysis of nonlinear partial differential equations with extreme learning machines
• Numerical solution of the nonlinear Klein-Gordon equation using radial basis functions
• On the approximation of functions by tanh neural networks
• On the convergence of physics informed neural networks for linear second-order elliptic and parabolic type PDEs
• Physics informed neural networks (PINNs) for approximating nonlinear dispersive PDEs
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Scientific machine learning through physics-informed neural networks: where we are and what's next
• The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
• VAE-KRnet and its applications to variational Bayes
• When and why PINNs fail to train: a neural tangent kernel perspective
• When do extended physics-informed neural networks (XPINNs) improve generalization?

Cited in

(22)

• Space-time error estimates for deep neural network approximations for differential equations
• A Rate of Convergence of Physics Informed Neural Networks for the Linear Second Order Elliptic PDEs
• Convergence of Physics-Informed Neural Networks Applied to Linear Second-Order Elliptic Interface Problems
• Time-dependent Dirac equation with physics-informed neural networks: computation and properties
• A decoupled physics-informed neural network for recovering a space-dependent force function in the wave equation from integral overdetermination data
• Higher-order error estimates for physics-informed neural networks approximating the primitive equations
• TDOR-MPINNs: multi-output physics-informed neural networks based on time differential order reduction for solving coupled Klein-Gordon-Zakharov systems
• Reconstruction and prediction of Volterra integral equations driven by Gaussian noise
• An adaptive residual-structured and scheduled loss weighting framework for enhancing physics-informed neural networks in solving partial differential equations
• HOMO-PINN: hyperparameter optimization of a multi-output physics-informed neural network
• Bound-preserving physics-informed neural networks for steady-state convection-diffusion-reaction problems
• A hybrid virtual element method and deep learning approach for solving one-dimensional Euler-Bernoulli beams
• Solving nonlinear Boussinesq equation of second-order time derivatives with physics-informed neural networks
• A Kolmogorov high order deep neural network for high frequency partial differential equations in high dimensions
• A nonoverlapping domain decomposition method for extreme learning machines: elliptic problems
• On the convergence of PINNs
• Error analysis and numerical algorithm for PDE approximation with hidden-layer concatenated physics informed neural networks
• Multiscale modeling of wave propagation with exponential integration and GMsFEM
• A new hybrid approach for solving partial differential equations: combining physics-informed neural networks with cat-and-mouse based optimization
• A deep learning approach: physics-informed neural networks for solving the 2D nonlinear sine-Gordon equation
• Error analyses of the single hidden layer neural network based on the extreme learning machine for solving differential equations
• Rank-inspired neural network for solving partial differential equations

This page was built for publication: Physics-informed neural networks for approximating dynamic (hyperbolic) PDEs of second order in time: error analysis and algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6087958)