A Study on a Dialogue Management Model for Multi-Party Dialogues
| Project/Area Number |
14580415
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | The University of Tokyo (2003)
Japan Advanced Institute of Science and Technology (2002) |
|---|
Principal Investigator |
ISHIZAKI Masato The University of Tokyo, The Institute for Socio Information and Communication Studies, Associate Professor, 社会情報研究所, 助教授 (30303340)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
SHIMAZU Akira Japan Advanced Institute of Science and Technology, School of Information Science, Professor, 情報科学研究科, 教授 (60293388)
|
|---|
| Project Period (FY) |
2002 – 2003
|
| Project Status |
Completed (Fiscal Year 2003)
|
| Budget Amount *help |
¥3,500,000 (Direct Cost: ¥3,500,000)
Fiscal Year 2003: ¥1,600,000 (Direct Cost: ¥1,600,000)
Fiscal Year 2002: ¥1,900,000 (Direct Cost: ¥1,900,000)
|
|---|
| Keywords | multi-party dialogues / chat dialogues / turn-taking system / adjacency pair / next position / resource / 話者交代システム / 意味的つながり / 話題スレッド / 表層的な手がかり / チャット対話システム / 擬似的重複 / 複数話題の同時進行 / 対話状況の共有 |
|---|
| Research Abstract |
Analysis of multi-party chat dialogues suggests that the turn-taking system might primarily concern the next position to the current utterance, not. the timing of the next utterance. The second part (next position) in response to the first part in the adjacency pair has special roles: The second part is assumed to be relevant to the first part; and the lack of the second part is considered to be a marked phenomenon.
In face-to-face dialogues, orientation to the timing of utterances produces utterances in the next position. However, in multi-party chat dialogues, the participants cannot control the timing of utterances. If the turn-taking system primarily concerns the timing of utterances in multi-party dialogues, the participants should make their turns shorter, which is not observed in the dialogue data. Rather they use a copy-and-paste technique and lexical items like nouns and proper nouns as a resource to display the next position to the target utterance, and, moreover, use special symbols to specify the next speaker and constrain the next utterance.
The finding here can account for why our newly built chat system, which enables users to share dialogue processes such as the timing of utterances, could not realize the smoothness of dialogue processes.
|
Report
(3 results)
Research Products
(19 results)
