Training Deep Surrogate Models with Large Scale Online Learning

TL;DR

This paper introduces an online training framework for deep surrogate models of PDEs, enabling larger datasets and improved generalization by simultaneous simulation and training, surpassing traditional offline methods.

Contribution

It proposes a scalable online training approach that reduces I/O bottlenecks, allowing training on larger datasets and enhancing model accuracy for PDE surrogate modeling.

Findings

Deep surrogate models trained online show increased dataset diversity.

Model accuracy improved by up to 68% with larger datasets.

Online training outperforms offline training in generalization capabilities.

Abstract

The spatiotemporal resolution of Partial Differential Equations (PDEs) plays important roles in the mathematical description of the world's physical phenomena. In general, scientists and engineers solve PDEs numerically by the use of computationally demanding solvers. Recently, deep learning algorithms have emerged as a viable alternative for obtaining fast solutions for PDEs. Models are usually trained on synthetic data generated by solvers, stored on disk and read back for training. This paper advocates that relying on a traditional static dataset to train these models does not allow the full benefit of the solver to be used as a data generator. It proposes an open source online training framework for deep surrogate models. The framework implements several levels of parallelism focused on simultaneously generating numerical simulations and training deep neural networks. This approach suppresses the I/O and storage bottleneck associated with disk-loaded datasets, and opens the way to training on significantly larger datasets. Experiments compare the offline and online training of four surrogate models, including state-of-the-art architectures. Results indicate that exposing deep surrogate models to more dataset diversity, up to hundreds of GB, can increase model generalization capabilities. Fully connected neural networks, Fourier Neural Operator (FNO), and Message Passing PDE Solver prediction accuracy is improved by 68%, 16% and 7%, respectively.