Learning Deep Representation with Energy-Based Self-Expressiveness for Subspace Clustering

TL;DR

This paper introduces a deep subspace clustering method using energy-based models that enables mini-batch training, allowing the use of deeper neural networks and self-supervised learning to improve clustering performance.

Contribution

It proposes an energy-based network for self-expressive coefficients, enabling mini-batch training and the integration of self-supervised learning for improved subspace clustering.

Findings

Significantly outperforms recent methods like SENet on multiple datasets.

Achieves 13.8% to 20.8% improvements in accuracy, NMI, and ARI.

Enables training of deeper neural networks for subspace clustering.

Abstract

Deep subspace clustering has attracted increasing attention in recent years. Almost all the existing works are required to load the whole training data into one batch for learning the self-expressive coefficients in the framework of deep learning. Although these methods achieve promising results, such a learning fashion severely prevents from the usage of deeper neural network architectures (e.g., ResNet), leading to the limited representation abilities of the models. In this paper, we propose a new deep subspace clustering framework, motivated by the energy-based models. In contrast to previous approaches taking the weights of a fully connected layer as the self-expressive coefficients, we propose to learn an energy-based network to obtain the self-expressive coefficients by mini-batch training. By this means, it is no longer necessary to load all data into one batch for learning, and it thus becomes a reality that we can utilize deeper neural network models for subspace clustering. Considering the powerful representation ability of the recently popular self-supervised learning, we attempt to leverage self-supervised representation learning to learn the dictionary. Finally, we propose a joint framework to learn both the self-expressive coefficients and dictionary simultaneously, and train the model in an end-to-end manner. The experiments are performed on three publicly available datasets, and extensive experimental results demonstrate our method can significantly outperform the other related approaches. For instance, on the three datasets, our method can averagely achieve $13.8\%$, $15.4\%$, $20.8\%$ improvements in terms of Accuracy, NMI, and ARI over SENet which is proposed very recently and obtains the second best results in the experiments.