LipnLab🚀 Excited to share that a survey paper from our RCLN team has been accepted at IJCAI 2026! This work has been done in collaboration with CentraleSupélec.
"Graph Rewiring in GNNs to Mitigate Over-Squashing and Over-Smoothing: A Survey" By Hugo Attali, Nathalie Pernelle, Davide Buscaldi, and Fragkiskos D. Malliaros.
📎 https://arxiv.org/abs/2411.17429
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.Our survey also opens a broader discussion on the limits and open questions of graph rewiring: when is modifying the topology truly necessary? How can observed improvements be properly attributed to connectivity changes rather than feature-driven effects? We argue that progress in this area will require clearer problem formulations, more explicit assumptions, and evaluation protocols that make results robust and comparable across settings positioning graph rewiring as a principled structural intervention to better understand how topology shapes learning in GNNs.
Looking forward to presenting our work at IJCAI 2026 at Bremen!
#GNN #MachineLearning #IJCAI2026 #GraphNeuralNetworks #AI #DeepLearning #Research #LIPN
Graph Rewiring in GNNs to Mitigate Over-Squashing and Over-Smoothing: A Survey
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.
艾天旅行
Grosses
SciPost Physics
