Unsupervised Visual Representation Learning by Online Constrained K-Means

TL;DR

This paper introduces CoKe, an online constrained K-Means method for unsupervised visual representation learning that effectively addresses clustering challenges and achieves competitive results with a single data view.

Contribution

It proposes a novel online constrained K-Means approach with theoretical guarantees, enabling efficient and effective unsupervised visual representation learning.

Findings

Achieves competitive performance on ImageNet and benchmarks.

Offers an online clustering method with theoretical optimality guarantees.

Demonstrates efficiency and efficacy in unsupervised learning tasks.

Abstract

Cluster discrimination is an effective pretext task for unsupervised representation learning, which often consists of two phases: clustering and discrimination. Clustering is to assign each instance a pseudo label that will be used to learn representations in discrimination. The main challenge resides in clustering since prevalent clustering methods (e.g., k-means) have to run in a batch mode. Besides, there can be a trivial solution consisting of a dominating cluster. To address these challenges, we first investigate the objective of clustering-based representation learning. Based on this, we propose a novel clustering-based pretext task with online \textbf{Co}nstrained \textbf{K}-m\textbf{e}ans (\textbf{CoKe}). Compared with the balanced clustering that each cluster has exactly the same size, we only constrain the minimal size of each cluster to flexibly capture the inherent data structure. More importantly, our online assignment method has a theoretical guarantee to approach the global optimum. By decoupling clustering and discrimination, CoKe can achieve competitive performance when optimizing with only a single view from each instance. Extensive experiments on ImageNet and other benchmark data sets verify both the efficacy and efficiency of our proposal. Code is available at \url{https://github.com/idstcv/CoKe}.