Paper: Controlling Listening-oriented Dialogue using Partially Observable Markov Decision Processes

ACL ID C10-1086
Title Controlling Listening-oriented Dialogue using Partially Observable Markov Decision Processes
Venue International Conference on Computational Linguistics
Session Main Conference
Year 2010
Authors

This paper investigates how to automat- ically create a dialogue control compo- nent of a listening agent to reduce the cur- rent high cost of manually creating such components. We collected a large number of listening-oriented dialogues with their user satisfaction ratings and used them to create adialoguecontrol componentusing partially observableMarkov decision pro- cesses (POMDPs), which can learn a pol- icy to satisfy users by automatically find- ing a reasonable reward function. A com- parison between our POMDP-based com- ponent and other similarly motivated sys- tems using human subjects revealed that POMDPs can satisfactorilyproduce a dia- logue control component that can achieve reasonable subjectiveassessment.