Generalization Error in the Deep Ritz Method with Smooth Activation Functions

Year:    2024

Author:    Janne Siipola

Communications in Computational Physics, Vol. 35 (2024), Iss. 3 : pp. 761–815

Abstract

Deep Ritz method is a deep learning paradigm to solve partial differential equations. In this article we study the generalization error of the Deep Ritz method. We focus on the quintessential problem which is the Poisson’s equation. We show that generalization error of the Deep Ritz method converges to zero with rate $\frac{C}{\sqrt{n}},$ and we discuss about the constant $C.$ Results are obtained for shallow and residual neural networks with smooth activation functions.

Journal Article Details

Publisher Name:    Global Science Press

Language:    English

DOI:    https://doi.org/10.4208/cicp.OA-2023-0253

Communications in Computational Physics, Vol. 35 (2024), Iss. 3 : pp. 761–815

Published online:    2024-01

AMS Subject Headings:    Global Science Press

Copyright:    COPYRIGHT: © Global Science Press

Pages:    55

Keywords:    Deep learning Deep Ritz method Poisson’s equation residual neural networks shallow neural networks generalization.

Author Details

Janne Siipola