Deep Operator Networks for Bayesian Parameter Estimation in PDEs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Raj, Amogh, Gudumotou, Carol Eunice, Bun, Sakol, Srinivasa, Keerthana, Sarshar, Arash
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912515430023168
author Raj, Amogh
Gudumotou, Carol Eunice
Bun, Sakol
Srinivasa, Keerthana
Sarshar, Arash
author_facet Raj, Amogh
Gudumotou, Carol Eunice
Bun, Sakol
Srinivasa, Keerthana
Sarshar, Arash
contents We present a novel framework combining Deep Operator Networks (DeepONets) with Physics-Informed Neural Networks (PINNs) to solve partial differential equations (PDEs) and estimate their unknown parameters. By integrating data-driven learning with physical constraints, our method achieves robust and accurate solutions across diverse scenarios. Bayesian training is implemented through variational inference, allowing for comprehensive uncertainty quantification for both aleatoric and epistemic uncertainties. This ensures reliable predictions and parameter estimates even in noisy conditions or when some of the physical equations governing the problem are missing. The framework demonstrates its efficacy in solving forward and inverse problems, including the 1D unsteady heat equation and 2D reaction-diffusion equations, as well as regression tasks with sparse, noisy observations. This approach provides a computationally efficient and generalizable method for addressing uncertainty quantification in PDE surrogate modeling.
format Preprint
id arxiv_https___arxiv_org_abs_2501_10684
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Deep Operator Networks for Bayesian Parameter Estimation in PDEs
Raj, Amogh
Gudumotou, Carol Eunice
Bun, Sakol
Srinivasa, Keerthana
Sarshar, Arash
Machine Learning
Computational Engineering, Finance, and Science
65K10, 35R30, 68T07
I.2.9; G.1.8; I.6.5
We present a novel framework combining Deep Operator Networks (DeepONets) with Physics-Informed Neural Networks (PINNs) to solve partial differential equations (PDEs) and estimate their unknown parameters. By integrating data-driven learning with physical constraints, our method achieves robust and accurate solutions across diverse scenarios. Bayesian training is implemented through variational inference, allowing for comprehensive uncertainty quantification for both aleatoric and epistemic uncertainties. This ensures reliable predictions and parameter estimates even in noisy conditions or when some of the physical equations governing the problem are missing. The framework demonstrates its efficacy in solving forward and inverse problems, including the 1D unsteady heat equation and 2D reaction-diffusion equations, as well as regression tasks with sparse, noisy observations. This approach provides a computationally efficient and generalizable method for addressing uncertainty quantification in PDE surrogate modeling.
title Deep Operator Networks for Bayesian Parameter Estimation in PDEs
topic Machine Learning
Computational Engineering, Finance, and Science
65K10, 35R30, 68T07
I.2.9; G.1.8; I.6.5
url https://arxiv.org/abs/2501.10684