We propose a novel end-to-end clustering training schedule for neural networks that is direct, i.e. the output is a probability distribution over cluster memberships. A neural network maps images to embeddings. We introduce centroid variables that have the same shape as image embeddings. These variables are jointly optimized with the network’s parameters. This is achieved by a cost function that associates the centroid variables with embeddings of input images. Finally, an additional layer maps embeddings to logits, allowing for the direct estimation of the respective cluster membership. Unlike other methods, this does not require any additional classifier to be trained on the embeddings in a separate step. The proposed approach achieves state-of-the-art results in unsupervised classification and we provide an extensive ablation study to demonstrate its capabilities.
CITATION STYLE
Haeusser, P., Plapp, J., Golkov, V., Aljalbout, E., & Cremers, D. (2019). Associative Deep Clustering: Training a Classification Network with No Labels. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11269 LNCS, pp. 18–32). Springer Verlag. https://doi.org/10.1007/978-3-030-12939-2_2
Mendeley helps you to discover research relevant for your work.