Dual-disentangled Deep Multiple Clustering

TL;DR

DDMC introduces a variational EM framework to address misalignment between learned representations and the goal of distinct clusterings in deep multiple clustering. It learns dual disentangled representations, combining coarse-grained augmentation-driven diversity with fine-grained latent factors, and couples them with a cluster-assignment objective across $M$ clusterings. The method derives an ELBO-based objective for fine-grained disentanglement, uses a KL capacity schedule, and alternates E-steps and M-steps to optimize both representation quality and clustering performance. Across seven benchmarks, DDMC achieves state-of-the-art results for both individual clusterings and the overall multi-clustering task, with ablations confirming the importance of coarse/fine disentanglement and the cluster-assignment component. Code is released at the provided project URL.

Abstract

Multiple clustering has gathered significant attention in recent years due to its potential to reveal multiple hidden structures of the data from different perspectives. Most of multiple clustering methods first derive feature representations by controlling the dissimilarity among them, subsequently employing traditional clustering methods (e.g., k-means) to achieve the final multiple clustering outcomes. However, the learned feature representations can exhibit a weak relevance to the ultimate goal of distinct clustering. Moreover, these features are often not explicitly learned for the purpose of clustering. Therefore, in this paper, we propose a novel Dual-Disentangled deep Multiple Clustering method named DDMC by learning disentangled representations. Specifically, DDMC is achieved by a variational Expectation-Maximization (EM) framework. In the E-step, the disentanglement learning module employs coarse-grained and fine-grained disentangled representations to obtain a more diverse set of latent factors from the data. In the M-step, the cluster assignment module utilizes a cluster objective function to augment the effectiveness of the cluster output. Our extensive experiments demonstrate that DDMC consistently outperforms state-of-the-art methods across seven commonly used tasks. Our code is available at https://github.com/Alexander-Yao/DDMC.

Figures (6)

• Figure 1: An example of multiple clustering that can reveal two or more distinct clusterings (i.e., $C^1$ for color and $C^2$ for shape).
• Figure 2: DDMC framework trains disentanglement learning and cluster assignment in an EM framework. During the E-step, the disentangled representation is learned through both coarse-grained and fine-grained disentangled representation learning. The learned disentangled representations can be applied to multiple clustering tasks. In the M-step, cluster assignment is optimized, enhancing the cluster-level performance.
• Figure 3: Visualization of DDMC and its variants color representations on the Fruit dataset.
• Figure 4: Results of parameter sensitivity of $K$.
• Figure 5: Results of parameter sensitivity of $T$.