Deep Clustering using Auto-Clustering Output Layer

In this paper, we propose a novel method to enrich the representation provided to the output layer of feedforward neural networks in the form of an auto-clustering output layer (ACOL) which enables the network to naturally create sub-clusters under the provided main class labels. In addition, a novel regularization term is introduced which allows ACOL to encourage the neural network to reveal its own explicit clustering objective. While the underlying process of finding the subclasses is completely unsupervised, semi-supervised learning is also possible based on the provided classification objective. The results show that ACOL can achieve a 99.2% clustering accuracy for the semi-supervised case when partial class labels are presented and a 96% accuracy for the unsupervised clustering case. These findings represent a paradigm shift especially when it comes to harnessing the power of deep networks for primary and secondary clustering applications in large datasets.