Saved in:
Bibliographic Details
Main Authors: Huang, Jianlei, Härkönen, Marc, Lange-Hegermann, Markus, Raiţă, Bogdan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.16663
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916972513460224
author Huang, Jianlei
Härkönen, Marc
Lange-Hegermann, Markus
Raiţă, Bogdan
author_facet Huang, Jianlei
Härkönen, Marc
Lange-Hegermann, Markus
Raiţă, Bogdan
contents Working with systems of partial differential equations (PDEs) is a fundamental task in computational science. Well-posed systems are addressed by numerical solvers or neural operators, whereas systems described by data are often addressed by PINNs or Gaussian processes. In this work, we propose Boundary Ehrenpreis--Palamodov Gaussian Processes (B-EPGPs), a novel probabilistic framework for constructing GP priors that satisfy both general systems of linear PDEs with constant coefficients and linear boundary conditions and can be conditioned on a finite data set. We explicitly construct GP priors for representative PDE systems with practical boundary conditions. Formal proofs of correctness are provided and empirical results demonstrating significant accuracy and computational resource improvements over state-of-the-art approaches.
format Preprint
id arxiv_https___arxiv_org_abs_2411_16663
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Gaussian Process Priors for Boundary Value Problems of Linear Partial Differential Equations
Huang, Jianlei
Härkönen, Marc
Lange-Hegermann, Markus
Raiţă, Bogdan
Machine Learning
Numerical Analysis
Commutative Algebra
60G15, 13N10, 13P25, 60-08, 35G35
Working with systems of partial differential equations (PDEs) is a fundamental task in computational science. Well-posed systems are addressed by numerical solvers or neural operators, whereas systems described by data are often addressed by PINNs or Gaussian processes. In this work, we propose Boundary Ehrenpreis--Palamodov Gaussian Processes (B-EPGPs), a novel probabilistic framework for constructing GP priors that satisfy both general systems of linear PDEs with constant coefficients and linear boundary conditions and can be conditioned on a finite data set. We explicitly construct GP priors for representative PDE systems with practical boundary conditions. Formal proofs of correctness are provided and empirical results demonstrating significant accuracy and computational resource improvements over state-of-the-art approaches.
title Gaussian Process Priors for Boundary Value Problems of Linear Partial Differential Equations
topic Machine Learning
Numerical Analysis
Commutative Algebra
60G15, 13N10, 13P25, 60-08, 35G35
url https://arxiv.org/abs/2411.16663