Improved training of physics-informed neural networks for parabolic differential equations with sharply perturbed initial conditions

DOI10.1016/j.cma.2023.116125arXiv2208.08635MaRDI QIDQ6171154

Alexandre M. Tartakovsky, Yifei Zong, Qizhi He

Publication date: 11 August 2023

Published in: Computer Methods in Applied Mechanics and Engineering (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2208.08635

zbMATH Keywords

importance samplinginverse problemsparabolic equationsphysics-informed neural networksbackward advection-dispersion equationsdeep neural network training

Mathematics Subject Classification ID

Numerical analysis (65-XX) Partial differential equations (35-XX)

Related Items (1)

Physics-informed machine learning method with space-time Karhunen-Loève expansions for forward and inverse partial differential equations

Cites Work

• Unnamed Item
• Unnamed Item
• On the limited memory BFGS method for large scale optimization
• Approximation and estimation bounds for artificial neural networks
• Reasoning, nonmonotonicity and learning in connectionist networks that capture propositional knowledge
• Exponential convergence of the deep neural network approximation for analytic functions
• DGM: a deep learning algorithm for solving partial differential equations
• A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics
• Physics-informed machine learning with conditional Karhunen-Loève expansions
• When and why PINNs fail to train: a neural tangent kernel perspective
• Scientific machine learning through physics-informed neural networks: where we are and what's next
• Physics-informed cokriging: a Gaussian-process-regression-based multifidelity method for data-model convergence
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• A comprehensive study of non-adaptive and residual-based adaptive sampling for physics-informed neural networks
• DAS-PINNs: a deep adaptive sampling method for solving high-dimensional partial differential equations
• Normalization methods for input and output vectors in backpropagation neural networks
• Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks
• Solving high-dimensional partial differential equations using deep learning
• On the Convergence of Physics Informed Neural Networks for Linear Second-Order Elliptic and Parabolic Type PDEs
• Adaptive importance sampling in monte carlo integration
• Failure-Informed Adaptive Sampling for PINNs

This page was built for publication: Improved training of physics-informed neural networks for parabolic differential equations with sharply perturbed initial conditions