Learning domain structure through probabilistic policy reuse in reinforcement learning.
Fernando FernándezManuela M. VelosoPublished in: Prog. Artif. Intell. (2013)
Keyphrases
- reinforcement learning
- learning process
- learning algorithm
- temporal difference
- actor critic
- supervised learning
- learned knowledge
- action selection
- temporal difference learning
- function approximators
- optimal policy
- partially observable
- function approximation
- complex domains
- domain independent
- learning systems
- online learning
- probabilistic model
- learning problems
- markov decision processes
- collaborative learning
- domain specific
- reinforcement learning methods
- bayes net
- policy search
- eligibility traces
- partially observable environments