Knowledge Revision for Reinforcement Learning with Abstract MDPs.
Kyriakos EfthymiadisDaniel KudenkoPublished in: AAMAS (2015)
Keyphrases
- reinforcement learning
- markov decision processes
- state space
- knowledge base
- higher level
- knowledge management
- knowledge based systems
- markov decision process
- function approximation
- domain knowledge
- knowledge acquisition
- optimal policy
- complex domains
- knowledge discovery
- background knowledge
- knowledge representation
- prior knowledge
- temporal difference
- learning process
- expert systems
- learning algorithm
- reward function
- reinforcement learning algorithms
- control problems
- action space
- approximate dynamic programming
- continuous state spaces
- continuous state and action spaces