Collocation-based Robust Variational Physics-Informed Neural Networks (CRVPINN)

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Łoś, Marcin, Służalec, Tomasz, Maczuga, Paweł, Vilkha, Askold, Uriarte, Carlos, Paszyński, Maciej
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929545468182528
author Łoś, Marcin
Służalec, Tomasz
Maczuga, Paweł
Vilkha, Askold
Uriarte, Carlos
Paszyński, Maciej
author_facet Łoś, Marcin
Służalec, Tomasz
Maczuga, Paweł
Vilkha, Askold
Uriarte, Carlos
Paszyński, Maciej
contents Physics-Informed Neural Networks (PINNs) have been successfully applied to solve Partial Differential Equations (PDEs). Their loss function is founded on a strong residual minimization scheme. Variational Physics-Informed Neural Networks (VPINNs) are their natural extension to weak variational settings. In this context, the recent work of Robust Variational Physics-Informed Neural Networks (RVPINNs) highlights the importance of conveniently translating the norms of the underlying continuum-level spaces to the discrete level. Otherwise, VPINNs might become unrobust, implying that residual minimization might be highly uncorrelated with a desired minimization of the error in the energy norm. However, applying this robustness to VPINNs typically entails dealing with the inverse of a Gram matrix, usually producing slow convergence speeds during training. In this work, we accelerate the implementation of RVPINN, establishing a LU factorization of sparse Gram matrix in a kind of point-collocation scheme with the same spirit as original PINNs. We call out method the Collocation-based Robust Variational Physics Informed Neural Networks (CRVPINN). We test our efficient CRVPINN algorithm on Laplace, advection-diffusion, and Stokes problems in two spatial dimensions.
format Preprint
id arxiv_https___arxiv_org_abs_2401_02300
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Collocation-based Robust Variational Physics-Informed Neural Networks (CRVPINN)
Łoś, Marcin
Służalec, Tomasz
Maczuga, Paweł
Vilkha, Askold
Uriarte, Carlos
Paszyński, Maciej
Machine Learning
Numerical Analysis
65M99, 68T07
G.1.8; I.2; I.m; G.1.10; J.2
Physics-Informed Neural Networks (PINNs) have been successfully applied to solve Partial Differential Equations (PDEs). Their loss function is founded on a strong residual minimization scheme. Variational Physics-Informed Neural Networks (VPINNs) are their natural extension to weak variational settings. In this context, the recent work of Robust Variational Physics-Informed Neural Networks (RVPINNs) highlights the importance of conveniently translating the norms of the underlying continuum-level spaces to the discrete level. Otherwise, VPINNs might become unrobust, implying that residual minimization might be highly uncorrelated with a desired minimization of the error in the energy norm. However, applying this robustness to VPINNs typically entails dealing with the inverse of a Gram matrix, usually producing slow convergence speeds during training. In this work, we accelerate the implementation of RVPINN, establishing a LU factorization of sparse Gram matrix in a kind of point-collocation scheme with the same spirit as original PINNs. We call out method the Collocation-based Robust Variational Physics Informed Neural Networks (CRVPINN). We test our efficient CRVPINN algorithm on Laplace, advection-diffusion, and Stokes problems in two spatial dimensions.
title Collocation-based Robust Variational Physics-Informed Neural Networks (CRVPINN)
topic Machine Learning
Numerical Analysis
65M99, 68T07
G.1.8; I.2; I.m; G.1.10; J.2
url https://arxiv.org/abs/2401.02300