Spatio-Temporal Graph Scattering Transform

TL;DR

The paper introduces a mathematically designed spatio-temporal graph scattering transform (ST-GST) that offers a theoretically interpretable, training-free alternative to graph neural networks, with improved stability and performance in limited data scenarios.

Contribution

It extends scattering transforms to the spatio-temporal domain with fixed, mathematically designed filters, enabling theoretical analysis and better performance with limited training data.

Findings

ST-GST outperforms trained graph neural networks by 35% accuracy on MSR Action3D.

Separable spatio-temporal graphs yield better and more efficient transforms than joint graphs.

Nonlinearity is essential for the empirical success of ST-GST.

Abstract

Although spatio-temporal graph neural networks have achieved great empirical success in handling multiple correlated time series, they may be impractical in some real-world scenarios due to a lack of sufficient high-quality training data. Furthermore, spatio-temporal graph neural networks lack theoretical interpretation. To address these issues, we put forth a novel mathematically designed framework to analyze spatio-temporal data. Our proposed spatio-temporal graph scattering transform (ST-GST) extends traditional scattering transforms to the spatio-temporal domain. It performs iterative applications of spatio-temporal graph wavelets and nonlinear activation functions, which can be viewed as a forward pass of spatio-temporal graph convolutional networks without training. Since all the filter coefficients in ST-GST are mathematically designed, it is promising for the real-world scenarios with limited training data, and also allows for a theoretical analysis, which shows that the proposed ST-GST is stable to small perturbations of input signals and structures. Finally, our experiments show that i) ST-GST outperforms spatio-temporal graph convolutional networks by an increase of 35% in accuracy for MSR Action3D dataset; ii) it is better and computationally more efficient to design the transform based on separable spatio-temporal graphs than the joint ones; and iii) the nonlinearity in ST-GST is critical to empirical performance.