Presentation 2014-10-24
Analysis of interview dialog for building user-profiling dialog system considering motivation of conversation
Yuya CHIBA, Akinori ITO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A dialog system was to obtain user's profile appropriately in order to provide a topic of dialog or recommend information adapted to the user. In this research, we assumed an interview-based user profiling system which asks the user about his/her personal information actively. In such a system, the system can obtain the detailed information if the user want to talk about the provided topic, but the system cannot obtain the beneficial information if the user does not interested in talking with that topic. A present paper analyse the interview dialog between humans for an initial study to estimate the user's motivation of conversation. As a result, the evaluators can judge the user's motivation of conversation of the relatively long dialog of the single topic in an accuracy from 70% to 80%. In addition, when they evaluate the subdivided dialog data, we observed some correlation between the judgements of the evaluators, although the concordance of the evaluation was decreased. Finally, it was indicated that several multi-modal information is efficient to estimate the user's motivation of the conversation, such as prosodic information, linguistic information, gesture, and gaze activity of the user.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) spoken dialog system / dialog analysis / user state / multi-modal information
Paper # HCS2014-67
Date of Issue

Conference Information
Committee HCS
Conference Date 2014/10/16(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Human Communication Science (HCS)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Analysis of interview dialog for building user-profiling dialog system considering motivation of conversation
Sub Title (in English)
Keyword(1) spoken dialog system
Keyword(2) dialog analysis
Keyword(3) user state
Keyword(4) multi-modal information
1st Author's Name Yuya CHIBA
1st Author's Affiliation School of Engineering, Tohoku University()
2nd Author's Name Akinori ITO
2nd Author's Affiliation School of Engineering, Tohoku University
Date 2014-10-24
Paper # HCS2014-67
Volume (vol) vol.114
Number (no) 273
Page pp.pp.-
#Pages 6
Date of Issue