Disentangled Generative Models for Robust Prediction of System Dynamics

TL;DR

This paper introduces a method using disentangled generative models to improve the robustness and out-of-distribution generalization of dynamical system predictions, especially in long-term and unseen domain scenarios.

Contribution

It proposes a disentanglement approach that separates domain parameters from dynamics in the latent space, enhancing model robustness and generalization.

Findings

Disentangled VAEs better adapt to unseen domain parameters.

Disentanglement improves long-term prediction accuracy.

Models show enhanced out-of-distribution performance in video sequences.

Abstract

Deep neural networks have become increasingly of interest in dynamical system prediction, but out-of-distribution generalization and long-term stability still remains challenging. In this work, we treat the domain parameters of dynamical systems as factors of variation of the data generating process. By leveraging ideas from supervised disentanglement and causal factorization, we aim to separate the domain parameters from the dynamics in the latent space of generative models. In our experiments we model dynamics both in phase space and in video sequences and conduct rigorous OOD evaluations. Results indicate that disentangled VAEs adapt better to domain parameters spaces that were not present in the training data. At the same time, disentanglement can improve the long-term and out-of-distribution predictions of state-of-the-art models in video sequences.