Learning Discrete Representations via Information Maximizing Self-Augmented Training

Open in new window