Maximizing Information Gain in Partially Observable Environments via Prediction Rewards.
Yash SatsangiSungsu LimShimon WhitesonFrans A. OliehoekMartha WhitePublished in: AAMAS (2020)
Keyphrases
- information gain
- partially observable environments
- chi squared
- text categorization
- reinforcement learning
- decision trees
- mutual information
- feature selection
- inverse reinforcement learning
- correlation coefficient
- reward function
- naive bayes
- reinforcement learning algorithms
- partially observable
- markov decision processes
- classification accuracy
- feature subset
- state space