You Never Cluster Alone

TL;DR

This paper introduces twin-contrast clustering (TCC), a novel method that enhances unsupervised clustering by incorporating cluster-level context into contrastive learning, leading to improved performance on benchmarks.

Contribution

The paper proposes a new cluster-level contrastive learning framework with end-to-end training, linking instance and cluster representations without alternating steps.

Findings

TCC outperforms state-of-the-art methods on benchmarks.

The method effectively encodes cluster context into contrastive learning.

End-to-end training simplifies the clustering process.

Abstract

Recent advances in self-supervised learning with instance-level contrastive objectives facilitate unsupervised clustering. However, a standalone datum is not perceiving the context of the holistic cluster, and may undergo sub-optimal assignment. In this paper, we extend the mainstream contrastive learning paradigm to a cluster-level scheme, where all the data subjected to the same cluster contribute to a unified representation that encodes the context of each data group. Contrastive learning with this representation then rewards the assignment of each datum. To implement this vision, we propose twin-contrast clustering (TCC). We define a set of categorical variables as clustering assignment confidence, which links the instance-level learning track with the cluster-level one. On one hand, with the corresponding assignment variables being the weight, a weighted aggregation along the data points implements the set representation of a cluster. We further propose heuristic cluster augmentation equivalents to enable cluster-level contrastive learning. On the other hand, we derive the evidence lower-bound of the instance-level contrastive objective with the assignments. By reparametrizing the assignment variables, TCC is trained end-to-end, requiring no alternating steps. Extensive experiments show that TCC outperforms the state-of-the-art on challenging benchmarks.