oversquashing

Code for the paper: "On the Bottleneck of Graph Neural Networks and Its Practical Implications"

This repository holds code and other relevant files for the Learning on Graphs 2022 (LoG) tutorial "Graph Rewiring: From Theory to Applications in Fairness"

🐼PANDA: Expanded Width-Aware Message Passing Beyond Rewiring, ICML 2024

Official repository for On Over-Squashing in Message Passing Neural Networks (ICML 2023)

Spectral Graph Pruning Against Over-squashing and Over-smoothing (NeurIPS 2024)

[ICLR 2026] Oversmoothing, “Oversquashing”, Heterophily, Long-Range, and more: Demystifying Common Beliefs in Graph Machine Learning

This project implements mechanisms to mitigate over-squashing, improving GNN performance on tasks requiring deep Networks.

In this work, we demonstrate that oversquashing is not limited to long-range tasks, but can also arise in short-range problems.

Channel-capacity-constrained width and depth estimation for graph neural networks.

Exploring and visualizing limitations of message-passing paradigm for GNNs. 📉