Fetching the paper…

Learning Discrete Representations via Information Maximizing Self-Augmented Training · Around