Agenda-Based User Simulation for Bootstrapping a POMDP Dialogue System

This paper investigates the problem of bootstrapping a statistical dialogue manager without access to training data and proposes a new probabilistic agenda-based method for simulating user behaviour. In experiments with a statistical POMDP dialogue system, the simulator was realistic enough to successfully test the prototype system and train a dialogue policy. An extensive study with human subjects showed that the learned policy was highly competitive, with task completion rates above 90%.

[1]  Alexander I. Rudnicky.  AN AGENDA-BASED DIALOG MANAGEMENT ARCHITECTURE FOR SPOKEN LANGUAGE SYSTEMS , 1999 .

[2]  Roberto Pieraccini,et al.  A stochastic model of human-machine interaction for learning dialog strategies , 2000, IEEE Trans. Speech Audio Process..

[3]  Kallirroi Georgila,et al.  EVALUATING EFFECTIVENESS AND PORTABILITY OF REINFORCEMENT LEARNED DIALOGUE STRATEGIES WITH REAL USERS: THE TALK TOWNINFO EVALUATION , 2006, 2006 IEEE Spoken Language Technology Workshop.

[4]  Steve J. Young,et al.  USING POMDPS FOR DIALOG MANAGEMENT , 2006, 2006 IEEE Spoken Language Technology Workshop.

[5]  Thierry Dutoit,et al.  A probabilistic framework for dialog simulation and optimal strategy learning , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[6]  Steve J. Young,et al.  A survey of statistical user simulation techniques for reinforcement-learning of dialogue management strategies , 2006, The Knowledge Engineering Review.

[7]  Steve J. Young,et al.  Partially observable Markov decision processes for spoken dialog systems , 2007, Comput. Speech Lang..

[8]  Hui Ye,et al.  The Hidden Information State Approach to Dialog Management , 2007, 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07.