Active Trajectory Estimation for Partially Observed Markov Decision Processes via Conditional Entropy.
Timothy L. MolloyGirish N. NairPublished in: ECC (2021)
Keyphrases
- markov decision processes
- partially observed
- conditional entropy
- information theory
- optimal policy
- transition matrices
- expected reward
- state space
- mutual information
- reinforcement learning
- finite state
- policy iteration
- decision theoretic planning
- dynamic programming
- infinite horizon
- risk sensitive
- planning under uncertainty
- action space
- average reward
- model based reinforcement learning
- semi markov decision processes
- information theoretic
- markov decision process
- factored mdps
- finite horizon
- partially observable
- reinforcement learning algorithms
- bayes error rate
- decision processes
- average cost
- image processing
- reachability analysis
- reward function
- model free
- search algorithm
- real time dynamic programming
- computer vision