Learning with Good Feature Representations in Bandits and in RL with a Generative Model.
Tor LattimoreCsaba SzepesváriGellért WeiszPublished in: ICML (2020)
Keyphrases
- feature representations
- generative model
- reinforcement learning
- feature representation
- prior knowledge
- learned models
- learning process
- image representation
- classifier ensemble
- probabilistic model
- learning algorithm
- low level features
- semi supervised
- image classification
- deep belief networks
- em algorithm
- supervised learning
- low level
- active learning
- feature extraction
- conditional random fields
- text classification
- decision trees