Deep Modularity Networks with Diversity-Preserving Regularization

TL;DR

This paper introduces DMoN-DPR, an improved graph clustering method that incorporates diversity-preserving regularizations to enhance feature separation, assignment confidence, and cluster interpretability, outperforming previous models on benchmark datasets.

Contribution

It proposes three novel regularization terms for Deep Modularity Networks to explicitly promote feature-space diversity and assignment confidence in graph clustering.

Findings

Significant improvement in clustering metrics on benchmark datasets

Enhanced feature-space separation and assignment confidence

Statistically significant results with p≤0.05

Abstract

Graph clustering plays a crucial role in graph representation learning but often faces challenges in achieving feature-space diversity. While Deep Modularity Networks (DMoN) leverage modularity maximization and collapse regularization to ensure structural separation, they lack explicit mechanisms for feature-space separation, assignment dispersion, and assignment-confidence control. We address this limitation by proposing Deep Modularity Networks with Diversity-Preserving Regularization (DMoN-DPR), which introduces three novel regularization terms: distance-based for inter-cluster separation, variance-based for per-cluster assignment dispersion, and an assignment-entropy penalty with a small positive weight, encouraging more confident assignments gradually. Our method significantly enhances label-based clustering metrics on feature-rich benchmark datasets (paired two-tailed t-test, $p\leq0.05$), demonstrating the effectiveness of incorporating diversity-preserving regularizations in creating meaningful and interpretable clusters.