A Latent space solver for PDE generalization

TL;DR

This paper introduces a hybrid latent space solver for PDEs that employs iterative inference and solution initialization, demonstrating strong generalization across various PDE conditions in engineering applications.

Contribution

It presents a novel latent space PDE solver that enhances generalization through an iterative inferencing strategy combined with solution initialization.

Findings

Successfully generalizes to multiple PDE conditions

Outperforms traditional methods in engineering case

Demonstrates robustness in diverse PDE scenarios

Abstract

In this work we propose a hybrid solver to solve partial differential equation (PDE)s in the latent space. The solver uses an iterative inferencing strategy combined with solution initialization to improve generalization of PDE solutions. The solver is tested on an engineering case and the results show that it can generalize well to several PDE conditions.