Balanced Multi-Relational Graph Clustering

TL;DR

This paper introduces BMGC, a novel clustering method for multi-relational graphs that dynamically addresses view imbalance through dominant view mining, leading to improved clustering performance on real-world and synthetic datasets.

Contribution

It proposes a new metric for structural disparity and a balanced clustering framework that effectively handles view imbalance in multi-relational graphs.

Findings

BMGC achieves state-of-the-art clustering performance.

The Aggregation Class Distance effectively quantifies graph disparities.

Dominant view mining enhances representation learning in multi-relational graphs.

Abstract

Multi-relational graph clustering has demonstrated remarkable success in uncovering underlying patterns in complex networks. Representative methods manage to align different views motivated by advances in contrastive learning. Our empirical study finds the pervasive presence of imbalance in real-world graphs, which is in principle contradictory to the motivation of alignment. In this paper, we first propose a novel metric, the Aggregation Class Distance, to empirically quantify structural disparities among different graphs. To address the challenge of view imbalance, we propose Balanced Multi-Relational Graph Clustering (BMGC), comprising unsupervised dominant view mining and dual signals guided representation learning. It dynamically mines the dominant view throughout the training process, synergistically improving clustering performance with representation learning. Theoretical analysis ensures the effectiveness of dominant view mining. Extensive experiments and in-depth analysis on real-world and synthetic datasets showcase that BMGC achieves state-of-the-art performance, underscoring its superiority in addressing the view imbalance inherent in multi-relational graphs. The source code and datasets are available at https://github.com/zxlearningdeep/BMGC.