Reconsidering Representation Alignment for Multi-view Clustering

TL;DR

This paper challenges the necessity of aligning view representations in multi-view clustering, proposing a simple baseline that avoids alignment and adding contrastive learning to improve performance significantly.

Contribution

It introduces a baseline model that omits representation alignment and enhances it with contrastive learning to better prioritize views and improve clustering results.

Findings

Avoiding representation alignment can yield better clustering performance.

Contrastive learning further improves the baseline, surpassing current state-of-the-art methods.

The proposed approach is simple, effective, and scalable across multiple datasets.

Abstract

Aligning distributions of view representations is a core component of today's state of the art models for deep multi-view clustering. However, we identify several drawbacks with na\"ively aligning representation distributions. We demonstrate that these drawbacks both lead to less separable clusters in the representation space, and inhibit the model's ability to prioritize views. Based on these observations, we develop a simple baseline model for deep multi-view clustering. Our baseline model avoids representation alignment altogether, while performing similar to, or better than, the current state of the art. We also expand our baseline model by adding a contrastive learning component. This introduces a selective alignment procedure that preserves the model's ability to prioritize views. Our experiments show that the contrastive learning component enhances the baseline model, improving on the current state of the art by a large margin on several datasets.