Spectral Theory for Edge Pruning in Asynchronous Recurrent Graph Neural Networks

TL;DR

This paper introduces a spectral theory-based dynamic edge pruning method for Asynchronous Recurrent Graph Neural Networks, aiming to improve efficiency while maintaining performance in complex, dynamic graph learning tasks.

Contribution

It proposes a novel spectral approach utilizing eigenvalues of the graph Laplacian for effective edge pruning in ARGNNs, enhancing computational efficiency.

Findings

Effective reduction of edges without performance loss

Spectral properties guide pruning decisions

Improved efficiency in dynamic graph learning

Abstract

Graph Neural Networks (GNNs) have emerged as a powerful tool for learning on graph-structured data, finding applications in numerous domains including social network analysis and molecular biology. Within this broad category, Asynchronous Recurrent Graph Neural Networks (ARGNNs) stand out for their ability to capture complex dependencies in dynamic graphs, resembling living organisms' intricate and adaptive nature. However, their complexity often leads to large and computationally expensive models. Therefore, pruning unnecessary edges becomes crucial for enhancing efficiency without significantly compromising performance. This paper presents a dynamic pruning method based on graph spectral theory, leveraging the imaginary component of the eigenvalues of the network graph's Laplacian.