Deflation-PINNs: Learning Multiple Solutions for PDEs and Landau-de Gennes

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Disarò, Sean, Maity, Ruma Rani, Bacho, Aras
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910090985996288
author Disarò, Sean
Maity, Ruma Rani
Bacho, Aras
author_facet Disarò, Sean
Maity, Ruma Rani
Bacho, Aras
contents Nonlinear Partial Differential Equations (PDEs) are ubiquitous in mathematical physics and engineering. Although Physics-Informed Neural Networks (PINNs) have emerged as a powerful tool for solving PDE problems, they typically struggle to identify multiple distinct solutions, since they are designed to find one solution at a time. To address this limitation, we introduce Deflation-PINNs, a novel framework that integrates a deflation loss with an architecture based on PINNs and Deep Operator Networks (DeepONets). By incorporating a deflation term into the loss function, our method systematically forces the Deflation-PINN to seek and converge upon distinct finitely many solution branches. We provide theoretical evidence on the convergence of our model and demonstrate the efficacy of Deflation-PINNs through numerical experiments on the Landau-de Gennes model of liquid crystals, a system renowned for its complex energy landscape and multiple equilibrium states. Our results show that Deflation-PINNs can successfully identify and characterize multiple distinct crystal structures.
format Preprint
id arxiv_https___arxiv_org_abs_2603_27936
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Deflation-PINNs: Learning Multiple Solutions for PDEs and Landau-de Gennes
Disarò, Sean
Maity, Ruma Rani
Bacho, Aras
Numerical Analysis
Machine Learning
Nonlinear Partial Differential Equations (PDEs) are ubiquitous in mathematical physics and engineering. Although Physics-Informed Neural Networks (PINNs) have emerged as a powerful tool for solving PDE problems, they typically struggle to identify multiple distinct solutions, since they are designed to find one solution at a time. To address this limitation, we introduce Deflation-PINNs, a novel framework that integrates a deflation loss with an architecture based on PINNs and Deep Operator Networks (DeepONets). By incorporating a deflation term into the loss function, our method systematically forces the Deflation-PINN to seek and converge upon distinct finitely many solution branches. We provide theoretical evidence on the convergence of our model and demonstrate the efficacy of Deflation-PINNs through numerical experiments on the Landau-de Gennes model of liquid crystals, a system renowned for its complex energy landscape and multiple equilibrium states. Our results show that Deflation-PINNs can successfully identify and characterize multiple distinct crystal structures.
title Deflation-PINNs: Learning Multiple Solutions for PDEs and Landau-de Gennes
topic Numerical Analysis
Machine Learning
url https://arxiv.org/abs/2603.27936