Graph Rewiring in GNNs to Mitigate Over-Squashing and Over-Smoothing: A Survey

2026-05-01

AI Summary

This study addresses the degradation of node representations in deep graph neural networks (GNNs) caused by over-squashing and over-smoothing. It presents a systematic survey of graph rewiring techniques, offering the first comprehensive taxonomy and mechanistic analysis of existing approaches. The work elucidates how these methods enhance information propagation by strategically reconstructing graph topology. Covering theoretical foundations, implementation strategies, and associated performance trade-offs, the paper critically evaluates the effectiveness and limitations of prominent rewiring methods. By synthesizing current knowledge, it provides both theoretical insights and practical guidance for GNN architecture design, while outlining promising directions for future research on deep and large-scale graphs.

Abstract

Graph Neural Networks are powerful models for learning from graph-structured data, yet their effectiveness is often limited by two critical challenges: over-squashing, where information from distant nodes is excessively compressed, and over-smoothing, where repeated propagation makes node representations indistinguishable. Both phenomena stem from the interaction between message passing and the input topology, ultimately degrading information flow and limiting the performance of GNNs. In this survey, we examine graph rewiring techniques, a class of methods designed to modify the graph topology to enhance information propagation in GNNs. We provide a comprehensive review of state-of-the-art rewiring approaches, delving into their theoretical underpinnings, practical implementations, and performance trade-offs.

Problem

• over-squashing
• over-smoothing
• graph neural networks
• information propagation
• graph topology

Innovation

• graph rewiring
• over-squashing
• over-smoothing
• graph neural networks
• information propagation