An information-theoretic approach to curiosity-driven reinforcement learning.
Susanne StillDoina PrecupPublished in: Theory Biosci. (2012)
Keyphrases
- reinforcement learning
- function approximation
- learning algorithm
- data driven
- machine learning
- state space
- reinforcement learning algorithms
- markov decision processes
- databases
- markov decision process
- robot control
- control problems
- model free
- optimal policy
- action selection
- supervised learning
- learning process
- data sets
- direct policy search