Presentation 2003/12/12
User Utterance Segmentation for Information-Seeking Dialogue
Ryoko TOKUHISA, Ryuta TERASHIMA, Toshihiro WAKITA, Kentaro INUI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) One of the old but unsolved issues in designing mixed-initiative spoken dialogue systems is how to decide the timing of system utterances. This paper addresses the issue restricting the target domain to information-seeking dialogues. We first claim that it is critically important for a system to decide when it should takes the dialogue initiative, and define the notion of information-requesting utterance segments (IRUSs), which gives useful clues for turn/initiative taking. We then report the present results of our experiments in which we took a machine learning-based approach achieving 81.1% accuracy in segmenting user utterances into IRUSs. We also examined how speech recognition errors influenced the performance of our model, finding that it worked tolerantly to at most 30% word recognition errors.
Keyword(in Japanese) (See Japanese page)
Keyword(in English)
Paper # NLC2003-91
Date of Issue

Conference Information
Committee NLC
Conference Date 2003/12/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Vice Chair

Paper Information
Registration To Natural Language Understanding and Models of Communication (NLC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) User Utterance Segmentation for Information-Seeking Dialogue
Sub Title (in English)
1st Author's Name Ryoko TOKUHISA
1st Author's Affiliation Toyota Central R&D Labs., INC.()
2nd Author's Name Ryuta TERASHIMA
2nd Author's Affiliation Toyota Central R&D Labs., INC.
3rd Author's Name Toshihiro WAKITA
3rd Author's Affiliation Toyota Central R&D Labs., INC.
4th Author's Name Kentaro INUI
4th Author's Affiliation Graduate School of Information Science Nara Institute of Science and Technology
Date 2003/12/12
Paper # NLC2003-91
Volume (vol) vol.103
Number (no) 518
Page pp.pp.-
#Pages 6
Date of Issue