Unlocking Out-of-Distribution Generalization in Dynamics through Physics-Guided Augmentation

TL;DR

This paper introduces SPARK, a physics-guided data augmentation method that improves out-of-distribution generalization in dynamical system modeling by integrating physical knowledge into training data and prediction models.

Contribution

The paper presents SPARK, a novel augmentation technique that combines autoencoders and physics-based interpolation to enhance model robustness in scarce and shifting data environments.

Findings

SPARK outperforms existing methods on diverse benchmarks.

It significantly improves out-of-distribution prediction accuracy.

The approach is effective in data-scarce regimes.

Abstract

In dynamical system modeling, traditional numerical methods are limited by high computational costs, while modern data-driven approaches struggle with data scarcity and distribution shifts. To address these fundamental limitations, we first propose SPARK, a physics-guided quantitative augmentation plugin. Specifically, SPARK utilizes a reconstruction autoencoder to integrate physical parameters into a physics-rich discrete state dictionary. This state dictionary then acts as a structured dictionary of physical states, enabling the creation of new, physically-plausible training samples via principled interpolation in the latent space. Further, for downstream prediction, these augmented representations are seamlessly integrated with a Fourier-enhanced Graph ODE, a combination designed to robustly model the enriched data distribution while capturing long-term temporal dependencies. Extensive experiments on diverse benchmarks demonstrate that SPARK significantly outperforms state-of-the-art baselines, particularly in challenging out-of-distribution scenarios and data-scarce regimes, proving the efficacy of our physics-guided augmentation paradigm.