Policy Improvement for POMDPs Using Normalized Importance Sampling
Christian R. SheltonPublished in: CoRR (2013)
Keyphrases
- importance sampling
- monte carlo
- partially observable markov decision processes
- optimal policy
- markov chain
- partially observable
- kalman filter
- policy gradient
- particle filter
- variance reduction
- rare events
- markov decision processes
- particle filtering
- reinforcement learning
- point based value iteration
- belief state
- dynamic programming
- approximate inference
- infinite horizon
- markov chain monte carlo
- similarity measure
- expected reward
- posterior distribution
- state space
- active learning
- search space
- feature selection
- computer vision