DPM: A Novel Training Method for Physics-Informed Neural Networks in Extrapolation

Authors

  • Jungeun Kim Yonsei University
  • Kookjin Lee Sandia National Lab
  • Dongeun Lee Texas A&M University-Commerce
  • Sheo Yon Jhin Yonsei University
  • Noseong Park Yonsei University

Keywords:

(Deep) Neural Network Algorithms

Abstract

We present a method for learning dynamics of complex physical processes described by time-dependent nonlinear partial differential equations (PDEs). Our particular interest lies in extrapolating solutions in time beyond the range of temporal domain used in training. Our choice for a baseline method is physics-informed neural network (PINN) because the method parameterizes not only the solutions, but also the equations that describe the dynamics of physical processes. We demonstrate that PINN performs poorly on extrapolation tasks in many benchmark problems. To address this, we propose a novel method for better training PINN and demonstrate that our newly enhanced PINNs can accurately extrapolate solutions in time. Our method shows up to 72% smaller errors than state-of-the-art methods in terms of the standard L2-norm metric.

Downloads

Published

2021-05-18

How to Cite

Kim, J., Lee, K., Lee, D., Jhin, S. Y., & Park, N. (2021). DPM: A Novel Training Method for Physics-Informed Neural Networks in Extrapolation. Proceedings of the AAAI Conference on Artificial Intelligence, 35(9), 8146-8154. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/16992

Issue

Section

AAAI Technical Track on Machine Learning II