Partially Observable Markov Decision Processes with Continuous Observations for Dialogue Management.
Jason D. WilliamsPascal PoupartSteve J. YoungPublished in: SIGDIAL Workshop (2005)
Keyphrases
- dialogue management
- partially observable markov decision processes
- partially observable markov decision process
- finite state
- belief state
- planning under uncertainty
- reinforcement learning
- state space
- dynamical systems
- markov decision processes
- dialogue system
- dynamic programming
- spoken dialogue systems
- decision problems
- decision theoretic
- optimal policy
- multi agent
- natural language generation
- language understanding
- partially observable
- planning problems
- infinite horizon
- natural language
- dynamic bayesian networks
- approximate solutions
- state variables
- human body
- markov chain
- general purpose