A probabilistic constrained clustering for transfer learning and image category discovery

A constrained clustering formulation has been proposed to perform transfer learning and image category discovery using deep learning . The core idea is to formulate a clustering objective with pairwise constraints that can be used to train a deep clustering network . The cluster assignments and their underlying feature representations are jointly optimized end-to-end . The proposed objective directly minimizes the negative log-likelihood of cluster assignment with respect to the pairwise constrained constraints, has no hyper-parameters, and demonstrates improved scalability and performance on both supervised learning and unsupervised transfer learning . In this work, we provide a novel clusterering formulation to address scalability issues of previous work .

Links: PDF - Abstract

Code :

None

Keywords : learning - clustering - constrained - transfer - objective -

Leave a Reply

Your email address will not be published. Required fields are marked *