Free-energy-based reinforcement learning in a partially observable environment.
Makoto OtsukaJunichiro YoshimotoKenji DoyaPublished in: ESANN (2010)
Keyphrases
- partially observable
- reinforcement learning
- state space
- markov decision processes
- partially observable domains
- partial observability
- partially observable markov decision process
- decision problems
- dynamical systems
- partially observable environments
- initially unknown
- hidden state
- partial observations
- markov decision problems
- action models
- reward function
- function approximation
- reinforcement learning algorithms
- belief state
- infinite horizon
- machine learning
- mobile robot
- dynamic programming
- learning agent
- partially observable markov decision processes
- optimal policy
- temporal difference
- domain specific
- learning capabilities
- fully observable
- transfer learning