ESMC: MLLM-Based Embedding Selection for Explainable Multiple Clustering

TL;DR

This paper introduces ESMC, a novel method that uses multi-modal large language models to enable user-driven, multi-clustering with improved accuracy by leveraging hidden state embeddings and pseudo-label learning.

Contribution

The work presents a new approach that utilizes MLLMs' hidden states for feature-specific clustering and introduces a lightweight clustering head with pseudo-label learning.

Findings

Achieves competitive clustering performance across datasets

Effectively incorporates user-defined criteria for clustering

Enhances accuracy with pseudo-label learning

Abstract

Typical deep clustering methods, while achieving notable progress, can only provide one clustering result per dataset. This limitation arises from their assumption of a fixed underlying data distribution, which may fail to meet user needs and provide unsatisfactory clustering outcomes. Our work investigates how multi-modal large language models (MLLMs) can be leveraged to achieve user-driven clustering, emphasizing their adaptability to user-specified semantic requirements. However, directly using MLLM output for clustering has risks for producing unstructured and generic image descriptions instead of feature-specific and concrete ones. To address these issues, our method first discovers that MLLMs' hidden states of text tokens are strongly related to the corresponding features, and leverages these embeddings to perform clusterings from any user-defined criteria. We also employ a lightweight clustering head augmented with pseudo-label learning, significantly enhancing clustering accuracy. Extensive experiments demonstrate its competitive performance on diverse datasets and metrics.