Third International Conference on Spoken Language Processing (ICSLP 94)

Yokohama, Japan
September 18-22, 1994

A User-Initiated Dialogue Model and its Implementation for Spontaneous Human-Computer Interaction

Hiroshi Kanazawa, Shigenobu Seto, Hideki Hashimoto, Hideaki Shinchi, Yoichi Takebayashi

Toshiba Corporation, Research & Development Center, Kawasaki, Japan

This paper describes a user-initiated dialogue model and its implementation on TOSBURG II, a real-time task-oriented spontaneous speech dialogue system, for achieving robust human-computer interaction. In order to make TOS-BURG II user-initiative and flexible, we designed it on a semantic interpretation and user-initiated dialogue management basis. To implement the user-initiated dialogue model on a robust user-unspecified system, we have introduced the ATN (Augmented Transition Network) and represented the model, including eight user states and 19 system states, within the ATN framework. Through dialogue experiments on unspecified users, TOSBURG II has shown to be highly robust and friendly. It is confirmed that our dialogue model facilitates user-initiated spontaneous human-computer interaction.

Full Paper

Bibliographic reference.  Kanazawa, Hiroshi / Seto, Shigenobu / Hashimoto, Hideki / Shinchi, Hideaki / Takebayashi, Yoichi (1994): "A user-initiated dialogue model and its implementation for spontaneous human-computer interaction", In ICSLP-1994, 111-114.