A real-time system for summarizing human-human spontaneous spoken dialogues

We have built a prototype automatic dialogue summarizer-a real-time system that automatically generates simple summaries of completely spontaneous human-human spoken dialogues without the machine interrupting the natural flow of conversation. Two dialogue participants (client and clerk) discuss conference room reservations in Japanese, and the system dynamically updates summaries of what rooms were reserved or canceled for what times and by whom. This paper describes the system's architecture, its component technologies and its performance. We discuss the robustness, efficiency and effectiveness of the system, and the use of a spontaneous dialogue corpus for development and testing.

[1]  J. Cleary,et al.  \self-organized Language Modeling for Speech Recognition". In , 1997 .

[2]  Douglas E. Appelt,et al.  FASTUS: A Cascaded Finite-State Transducer for Extracting Information from Natural-Language Text , 1997, ArXiv.

[3]  Megumi Kameyama,et al.  Coping with aboutness complexity in information extraction from spoken dialogues , 1994, ICSLP.

[4]  Mitch Weintraub,et al.  The decipher speech recognition system , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[5]  Andreas Stolcke,et al.  Statistical language modeling for speech disfluencies , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.