Twenty Questions for Localizing Multiple Objects by Counting: Bayes Optimal Policies for Entropy Loss.
Weidong HanPeter I. FrazierBruno M. JedynakPublished in: CoRR (2014)
Keyphrases
- multiple objects
- optimal policy
- markov decision processes
- decision problems
- state space
- multi object
- reinforcement learning
- multiple object tracking
- dynamic programming
- multistage
- finite horizon
- particle filter
- average reward
- multiple images
- complex scenes
- finite state
- infinite horizon
- data association
- sufficient conditions
- long run
- markov decision process
- occlusion handling
- serial inventory systems
- multiple targets
- tracking of multiple objects
- initial state
- policy iteration
- dynamic programming algorithms
- search algorithm
- average cost
- average reward reinforcement learning
- occluded objects
- viewpoint
- semi markov decision processes
- lost sales
- inventory level
- three dimensional