Skip to Content

Graph Clustering with Graph Neural Networks

Deep Modularity Networks model performance
Our model outperforms different pooling methods and other baselines.

paper · pdf · code

TL;DR

End-to-end graph clsutering with graph neural networks.

In this paper:

  • We introduce an end-to-end unsupervised clustering module for GNNs.
  • We make a thorough empirical study of performance on synthetic graphs in various regimes.
  • We show that on real graphs we obtain state-of-the-art performance.