Controlling Listening-oriented Dialogue using Partially Observable Markov Decision Processes.
Toyomi MeguroRyuichiro HigashinakaYasuhiro MinamiKohji DohsakaPublished in: COLING (2010)
Keyphrases
- partially observable markov decision processes
- finite state
- dynamical systems
- belief state
- planning under uncertainty
- belief space
- continuous state
- reinforcement learning
- decision problems
- optimal policy
- partial observability
- partially observable stochastic games
- stochastic domains
- markov decision processes
- dynamic programming
- multi agent
- partially observable markov
- state space
- partially observable domains
- sequential decision making problems
- planning problems
- partially observable
- markov chain
- predictive state representations
- approximate solutions
- initial state
- image segmentation
- linear programming
- nash equilibrium
- belief revision