Leveraging Domain Knowledge for Reinforcement Learning Using MMC Architectures.
Rajkumar RamamurthyChristian BauckhageRafet SifaJannis SchückerStefan WrobelPublished in: ICANN (2) (2019)
Keyphrases
- domain knowledge
- reinforcement learning
- function approximation
- background knowledge
- reinforcement learning algorithms
- temporal difference
- learned knowledge
- previously learned
- markov decision processes
- domain experts
- multi agent reinforcement learning
- model free
- knowledge sources
- temporal difference learning
- multi class
- state space
- optimal policy
- learning agents
- domain ontology
- stochastic approximation
- domain theory
- learning process
- multi agent
- learning problems
- domain specific knowledge
- control problems
- supervised learning
- dynamic programming
- robot control
- case study
- multiagent systems
- domain models
- prior knowledge
- autonomous learning
- training data