Learning to Discover Novel Visual Categories via Deep Transfer Clustering

TL;DR

This paper introduces a transfer learning approach to discover new visual categories in unlabeled images by extending deep clustering techniques, leveraging prior knowledge, and estimating class numbers, achieving superior results on multiple benchmarks.

Contribution

It extends Deep Embedded Clustering to transfer learning, introduces a representation bottleneck, temporal ensembling, and a method for estimating class numbers using prior knowledge.

Findings

Outperforms state-of-the-art on ImageNet, CIFAR, SVHN, OmniGlot

Effectively estimates number of classes in unlabeled data

Improves clustering quality with transfer learning techniques

Abstract

We consider the problem of discovering novel object categories in an image collection. While these images are unlabelled, we also assume prior knowledge of related but different image classes. We use such prior knowledge to reduce the ambiguity of clustering, and improve the quality of the newly discovered classes. Our contributions are twofold. The first contribution is to extend Deep Embedded Clustering to a transfer learning setting; we also improve the algorithm by introducing a representation bottleneck, temporal ensembling, and consistency. The second contribution is a method to estimate the number of classes in the unlabelled data. This also transfers knowledge from the known classes, using them as probes to diagnose different choices for the number of classes in the unlabelled subset. We thoroughly evaluate our method, substantially outperforming state-of-the-art techniques in a large number of benchmarks, including ImageNet, OmniGlot, CIFAR-100, CIFAR-10, and SVHN.