Learning Diverse and Discriminative Representations via the Principle of Maximal Coding Rate Reduction.
Yaodong YuKwan Ho Ryan ChanChong YouChaobing SongYi MaPublished in: NeurIPS (2020)
Keyphrases
- learning algorithm
- reinforcement learning
- unsupervised learning
- online learning
- active learning
- prior knowledge
- learning process
- supervised learning
- object recognition
- real world
- learning systems
- learning tasks
- knowledge acquisition
- learning environment
- inductive inference
- weakly supervised
- structured prediction
- structured output
- positive training examples