Top-Down Deep Clustering with Multi-generator GANs

TL;DR

This paper introduces HC-MGAN, a novel deep clustering method using multi-generator GANs within a hierarchical top-down framework, achieving competitive results and meaningful semantic data organization.

Contribution

It pioneers the use of multi-generator GANs for hierarchical deep clustering, creating a top-down clustering tree that captures semantic data structures.

Findings

Achieves competitive clustering performance on benchmark datasets.

Provides a hierarchical clustering tree with semantically coherent patterns.

Demonstrates the effectiveness of multi-generator GANs in deep clustering.

Abstract

Deep clustering (DC) leverages the representation power of deep architectures to learn embedding spaces that are optimal for cluster analysis. This approach filters out low-level information irrelevant for clustering and has proven remarkably successful for high dimensional data spaces. Some DC methods employ Generative Adversarial Networks (GANs), motivated by the powerful latent representations these models are able to learn implicitly. In this work, we propose HC-MGAN, a new technique based on GANs with multiple generators (MGANs), which have not been explored for clustering. Our method is inspired by the observation that each generator of a MGAN tends to generate data that correlates with a sub-region of the real data distribution. We use this clustered generation to train a classifier for inferring from which generator a given image came from, thus providing a semantically meaningful clustering for the real distribution. Additionally, we design our method so that it is performed in a top-down hierarchical clustering tree, thus proposing the first hierarchical DC method, to the best of our knowledge. We conduct several experiments to evaluate the proposed method against recent DC methods, obtaining competitive results. Last, we perform an exploratory analysis of the hierarchical clustering tree that highlights how accurately it organizes the data in a hierarchy of semantically coherent patterns.