Unsupervised Community Detection with Modularity-Based Attention Model
This work addresses the problem of community detection in graphs for researchers and practitioners, offering an incremental improvement by combining attention models with modularity optimization.
The paper tackles unsupervised node clustering on graphs by proposing an algorithm that encodes Bethe Hessian embeddings with a soft modularity loss, achieving competitive performance with classical and Graph Neural Network models while being trainable on a single graph.
In this paper we take a problem of unsupervised nodes clustering on graphs and show how recent advances in attention models can be applied successfully in a "hard" regime of the problem. We propose an unsupervised algorithm that encodes Bethe Hessian embeddings by optimizing soft modularity loss and argue that our model is competitive to both classical and Graph Neural Network (GNN) models while it can be trained on a single graph.