PINN-DG: Residual neural network methods trained with Finite Elements

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Grekas, Georgios, Makridakis, Charalambos G., Pryer, Tristan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916833804681216
author Grekas, Georgios
Makridakis, Charalambos G.
Pryer, Tristan
author_facet Grekas, Georgios
Makridakis, Charalambos G.
Pryer, Tristan
contents Over the past few years, neural network methods have evolved in various directions for approximating partial differential equations (PDEs). A promising new development is the integration of neural networks with classical numerical techniques such as finite elements and finite differences. In this paper, we introduce a new class of Physics-Informed Neural Networks (PINNs) trained using discontinuous Galerkin finite element methods. Unlike standard collocation-based PINNs that rely on pointwise gradient evaluations and Monte Carlo quadrature, our approach computes the loss functional using finite element interpolation and integration. This avoids costly pointwise derivative computations, particularly advantageous for elliptic PDEs requiring second-order derivatives, and inherits key stability and accuracy benefits from the finite element framework. We present a convergence analysis based on variational arguments and support our theoretical findings with numerical experiments that demonstrate improved efficiency and robustness.
format Preprint
id arxiv_https___arxiv_org_abs_2507_03521
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PINN-DG: Residual neural network methods trained with Finite Elements
Grekas, Georgios
Makridakis, Charalambos G.
Pryer, Tristan
Numerical Analysis
Computational Physics
65M15 (Primary), 65M12 (Secondary)
Over the past few years, neural network methods have evolved in various directions for approximating partial differential equations (PDEs). A promising new development is the integration of neural networks with classical numerical techniques such as finite elements and finite differences. In this paper, we introduce a new class of Physics-Informed Neural Networks (PINNs) trained using discontinuous Galerkin finite element methods. Unlike standard collocation-based PINNs that rely on pointwise gradient evaluations and Monte Carlo quadrature, our approach computes the loss functional using finite element interpolation and integration. This avoids costly pointwise derivative computations, particularly advantageous for elliptic PDEs requiring second-order derivatives, and inherits key stability and accuracy benefits from the finite element framework. We present a convergence analysis based on variational arguments and support our theoretical findings with numerical experiments that demonstrate improved efficiency and robustness.
title PINN-DG: Residual neural network methods trained with Finite Elements
topic Numerical Analysis
Computational Physics
65M15 (Primary), 65M12 (Secondary)
url https://arxiv.org/abs/2507.03521