Optimally weighted loss functions for solving PDEs with Neural Networks

DOI10.1016/J.CAM.2021.113887MaRDI QIDQ6334842arXivWikidataFDO

Authors Remco van der Meer, Cornelis W. Oosterlee, Anastasia Borovykh

Publication date 14 February 2020

Full work available at URL https://doi.org/10.1016/J.CAM.2021.113887

Second-order elliptic equations (35J15) Probabilistic models, generic numerical methods in probability and statistics (65C20) Numerical optimization and variational techniques (65K10) Spectral, collocation and related methods for boundary value problems involving PDEs (65N35) Probabilistic methods, particle methods, etc. for boundary value problems involving PDEs (65N75) Numerical methods for partial differential equations, boundary value problems (65N99) Artificial neural networks and deep learning (68T07) Diffusion (76R50)

arXiv classification

cs.NA math.NA

Abstract: Recent works have shown that deep neural networks can be employed to solve partial differential equations, giving rise to the framework of physics informed neural networks. We introduce a generalization for these methods that manifests as a scaling parameter which balances the relative importance of the different constraints imposed by partial differential equations. A mathematical motivation of these generalized methods is provided, which shows that for linear and well-posed partial differential equations, the functional form is convex. We then derive a choice for the scaling parameter that is optimal with respect to a measure of relative error. Because this optimal choice relies on having full knowledge of analytical solutions, we also propose a heuristic method to approximate this optimal choice. The proposed methods are compared numerically to the original methods on a variety of model partial differential equations, with the number of data points being updated adaptively. For several problems, including high-dimensional PDEs the proposed methods are shown to significantly enhance accuracy.

This page was built for publication: Optimally weighted loss functions for solving PDEs with Neural Networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6334842)