A Bayesian Sampling Approach to Exploration in Reinforcement Learning.
John AsmuthLihong LiMichael L. LittmanAli NouriDavid WingatePublished in: UAI (2009)
Keyphrases
- reinforcement learning
- exploration exploitation
- active learning
- active exploration
- exploration strategy
- action selection
- bandit problems
- state space
- model based reinforcement learning
- sequential monte carlo
- function approximation
- model free
- posterior probability
- markov decision processes
- multi agent
- learning algorithm
- markov chain monte carlo
- bayesian networks
- multi armed bandit
- supervised learning
- policy search
- autonomous learning
- multi agent systems
- reinforcement learning algorithms
- data driven
- gaussian processes
- random sampling
- robotic control
- learning process
- bayesian decision
- sampling strategy
- proposal distribution
- multi agent reinforcement learning
- sampled data
- bayesian estimation
- sampling algorithm
- bayesian learning
- machine learning
- maximum likelihood
- optimal policy
- learning problems