Stable Cluster Discrimination for Deep Clustering

TL;DR

This paper introduces Stable Cluster Discrimination (SeCu), a novel deep clustering method that improves stability and performance by addressing the instability of discrimination tasks in one-stage clustering, validated on benchmarks and ImageNet.

Contribution

The paper proposes SeCu, a new stable cluster discrimination task with a hardness-aware criterion and global entropy constraint, enhancing one-stage deep clustering performance.

Findings

SeCu achieves state-of-the-art results on benchmark datasets.

SeCu effectively mitigates instability in one-stage clustering.

The method demonstrates strong performance on ImageNet.

Abstract

Deep clustering can optimize representations of instances (i.e., representation learning) and explore the inherent data distribution (i.e., clustering) simultaneously, which demonstrates a superior performance over conventional clustering methods with given features. However, the coupled objective implies a trivial solution that all instances collapse to the uniform features. To tackle the challenge, a two-stage training strategy is developed for decoupling, where it introduces an additional pre-training stage for representation learning and then fine-tunes the obtained model for clustering. Meanwhile, one-stage methods are developed mainly for representation learning rather than clustering, where various constraints for cluster assignments are designed to avoid collapsing explicitly. Despite the success of these methods, an appropriate learning objective tailored for deep clustering has not been investigated sufficiently. In this work, we first show that the prevalent discrimination task in supervised learning is unstable for one-stage clustering due to the lack of ground-truth labels and positive instances for certain clusters in each mini-batch. To mitigate the issue, a novel stable cluster discrimination (SeCu) task is proposed and a new hardness-aware clustering criterion can be obtained accordingly. Moreover, a global entropy constraint for cluster assignments is studied with efficient optimization. Extensive experiments are conducted on benchmark data sets and ImageNet. SeCu achieves state-of-the-art performance on all of them, which demonstrates the effectiveness of one-stage deep clustering. Code is available at \url{https://github.com/idstcv/SeCu}.