Saved in:
Bibliographic Details
Main Authors: De Ryck, Tim, Bonnet, Florent, Mishra, Siddhartha, de Bézenac, Emmanuel
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2310.05801
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913339758608384
author De Ryck, Tim
Bonnet, Florent
Mishra, Siddhartha
de Bézenac, Emmanuel
author_facet De Ryck, Tim
Bonnet, Florent
Mishra, Siddhartha
de Bézenac, Emmanuel
contents In this paper, we investigate the behavior of gradient descent algorithms in physics-informed machine learning methods like PINNs, which minimize residuals connected to partial differential equations (PDEs). Our key result is that the difficulty in training these models is closely related to the conditioning of a specific differential operator. This operator, in turn, is associated to the Hermitian square of the differential operator of the underlying PDE. If this operator is ill-conditioned, it results in slow or infeasible training. Therefore, preconditioning this operator is crucial. We employ both rigorous mathematical analysis and empirical evaluations to investigate various strategies, explaining how they better condition this critical operator, and consequently improve training.
format Preprint
id arxiv_https___arxiv_org_abs_2310_05801
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle An operator preconditioning perspective on training in physics-informed machine learning
De Ryck, Tim
Bonnet, Florent
Mishra, Siddhartha
de Bézenac, Emmanuel
Machine Learning
In this paper, we investigate the behavior of gradient descent algorithms in physics-informed machine learning methods like PINNs, which minimize residuals connected to partial differential equations (PDEs). Our key result is that the difficulty in training these models is closely related to the conditioning of a specific differential operator. This operator, in turn, is associated to the Hermitian square of the differential operator of the underlying PDE. If this operator is ill-conditioned, it results in slow or infeasible training. Therefore, preconditioning this operator is crucial. We employ both rigorous mathematical analysis and empirical evaluations to investigate various strategies, explaining how they better condition this critical operator, and consequently improve training.
title An operator preconditioning perspective on training in physics-informed machine learning
topic Machine Learning
url https://arxiv.org/abs/2310.05801