论文信息 - Acquiring and Evaluating a Dialog Corpus through a Dialog Simulation Technique

Acquiring and Evaluating a Dialog Corpus through a Dialog Simulation Technique

In this paper, we present an approach for automatically acquiring a dialog corpus by means of the interaction of a dialog manager and a user simulator. A random selection of the answers has been used for the operation of both modules, defining stop conditions for automatically deciding if the dialog is successful or not. Therefore, an initial corpus is not necessary to develop these two modules. In this work, we use a statistical dialog manager to evaluate the behavior of the corpus acquired using this approach. This dialog manager has been learned from the simulated corpus and has been evaluated using a previous corpus acquired for the task with real users.

David Griol | Encarna Segarra | Emilio Sanchis Arnal | Lluís F. Hurtado

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

[2] Encarna Segarra,et al. Development of a stochastic dialog manager driven by semantics , 2003, INTERSPEECH.

[3] Eduardo Lleida,et al. Design and acquisition of a telephone spontaneous speech dialogue corpus in Spanish: DIHANA , 2006, LREC.

[4] Steve Young,et al. The statistical approach to the design of spoken dialogue systems , 2003 .

[5] 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.

[6] David Griol,et al. A stochastic approach for dialog management based on neural networks , 2006, INTERSPEECH.

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