Joint Unsupervised Learning of Deep Representations and Image Clusters

TL;DR

This paper introduces a recurrent framework that jointly learns deep representations and image clusters in an unsupervised manner, improving clustering accuracy and representation quality through end-to-end optimization.

Contribution

It presents a novel recurrent framework that unifies clustering and representation learning, outperforming existing methods on multiple image datasets.

Findings

Outperforms state-of-the-art clustering methods

Produces representations that transfer well to other tasks

Effectively integrates clustering and deep learning in a single model

Abstract

In this paper, we propose a recurrent framework for Joint Unsupervised LEarning (JULE) of deep representations and image clusters. In our framework, successive operations in a clustering algorithm are expressed as steps in a recurrent process, stacked on top of representations output by a Convolutional Neural Network (CNN). During training, image clusters and representations are updated jointly: image clustering is conducted in the forward pass, while representation learning in the backward pass. Our key idea behind this framework is that good representations are beneficial to image clustering and clustering results provide supervisory signals to representation learning. By integrating two processes into a single model with a unified weighted triplet loss and optimizing it end-to-end, we can obtain not only more powerful representations, but also more precise image clusters. Extensive experiments show that our method outperforms the state-of-the-art on image clustering across a variety of image datasets. Moreover, the learned representations generalize well when transferred to other tasks.