Convolutional Autoencoders for Reduced-Order Modeling

TL;DR

This paper introduces convolutional autoencoders for nonlinear dimension reduction in reduced-order modeling of dynamical systems, enabling data-efficient training and application to multiple equations.

Contribution

It presents a novel approach using randomized training data and autoencoders for nonlinear reduction, independent of full-order model samples, and integrates manifold projection methods.

Findings

Effective nonlinear dimension reduction for wave and Kuramoto-Shivasinsky equations.

Training methods that do not rely on full-order model samples.

Successful application to heat, wave, and Kuramoto-Shivasinsky equations.

Abstract

In the construction of reduced-order models for dynamical systems, linear projection methods, such as proper orthogonal decompositions, are commonly employed. However, for many dynamical systems, the lower dimensional representation of the state space can most accurately be described by a \textit{nonlinear} manifold. Previous research has shown that deep learning can provide an efficient method for performing nonlinear dimension reduction, though they are dependent on the availability of training data and are often problem-specific \citep[see][]{carlberg_ca}. Here, we utilize randomized training data to create and train convolutional autoencoders that perform nonlinear dimension reduction for the wave and Kuramoto-Shivasinsky equations. Moreover, we present training methods that are independent of full-order model samples and use the manifold least-squares Petrov-Galerkin projection method to define a reduced-order model for the heat, wave, and Kuramoto-Shivasinsky equations using the same autoencoder.