Learnable Spectral Wavelets on Dynamic Graphs to Capture Global Interactions

TL;DR

This paper introduces a novel method for dynamic graph learning by capturing evolving spectra through learnable spectral wavelets, effectively integrating global and local features to improve performance on various tasks.

Contribution

The paper proposes a new approach to learn spectral wavelets on dynamic graphs, capturing global interactions over time, which enhances representation learning beyond local neighborhood aggregation.

Findings

Significantly outperforms existing methods on eight datasets

Effectively captures both local and global interactions

Improves dynamic graph learning tasks

Abstract

Learning on evolving(dynamic) graphs has caught the attention of researchers as static methods exhibit limited performance in this setting. The existing methods for dynamic graphs learn spatial features by local neighborhood aggregation, which essentially only captures the low pass signals and local interactions. In this work, we go beyond current approaches to incorporate global features for effectively learning representations of a dynamically evolving graph. We propose to do so by capturing the spectrum of the dynamic graph. Since static methods to learn the graph spectrum would not consider the history of the evolution of the spectrum as the graph evolves with time, we propose a novel approach to learn the graph wavelets to capture this evolving spectra. Further, we propose a framework that integrates the dynamically captured spectra in the form of these learnable wavelets into spatial features for incorporating local and global interactions. Experiments on eight standard datasets show that our method significantly outperforms related methods on various tasks for dynamic graphs.