Graph Contrastive Clustering

TL;DR

This paper introduces Graph Contrastive Clustering (GCC), a novel framework that leverages graph-based contrastive learning to improve clustering performance by incorporating category information and enhancing feature discriminability.

Contribution

The paper proposes a new graph contrastive learning framework for clustering that lifts instance-level to cluster-level consistency and integrates category information.

Findings

Outperforms state-of-the-art methods on six datasets.

Uses graph Laplacian-based contrastive loss for discriminative features.

Employs a graph-based strategy for compact clustering assignments.

Abstract

Recently, some contrastive learning methods have been proposed to simultaneously learn representations and clustering assignments, achieving significant improvements. However, these methods do not take the category information and clustering objective into consideration, thus the learned representations are not optimal for clustering and the performance might be limited. Towards this issue, we first propose a novel graph contrastive learning framework, which is then applied to the clustering task and we come up with the Graph Constrastive Clustering~(GCC) method. Different from basic contrastive clustering that only assumes an image and its augmentation should share similar representation and clustering assignments, we lift the instance-level consistency to the cluster-level consistency with the assumption that samples in one cluster and their augmentations should all be similar. Specifically, on the one hand, the graph Laplacian based contrastive loss is proposed to learn more discriminative and clustering-friendly features. On the other hand, a novel graph-based contrastive learning strategy is proposed to learn more compact clustering assignments. Both of them incorporate the latent category information to reduce the intra-cluster variance while increasing the inter-cluster variance. Experiments on six commonly used datasets demonstrate the superiority of our proposed approach over the state-of-the-art methods.