| Item type |
SIG Technical Reports(1) |
| 公開日 |
2016-09-22 |
| タイトル |
|
|
タイトル |
Evaluating Helpdesk Dialogues: Initial Considerations from An Information Access Perspective |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Evaluating Helpdesk Dialogues: Initial Considerations from An Information Access Perspective |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
評価と分析 |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
Waseda University |
| 著者所属 |
|
|
|
Waseda University |
| 著者所属 |
|
|
|
Tsinghua University／Waseda University |
| 著者所属(英) |
|
|
|
en |
|
|
Waseda University |
| 著者所属(英) |
|
|
|
en |
|
|
Waseda University |
| 著者所属(英) |
|
|
|
en |
|
|
Tsinghua University / Waseda University |
| 著者名 |
Tetsuya, Sakai
Zhaohao, Zeng
Cheng, Luo
|
| 著者名(英) |
Tetsuya, Sakai
Zhaohao, Zeng
Cheng, Luo
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Whenever a user of a commercial product or a service encounters a problem, an effective and efficient way to solve it is to contact the helpdesk. In the present study, we address the problem of evaluating textual dialogues between the customer and the helpdesk, such as those that take the form of online chats. Thus, we consider textual, dyadic and task-oriented dialogues. Successful dialogues are desirable both for the customer and for the company that sells the product / service. We are currently building a small-scale Chinese helpdesk dialogue test collection (with English translations) and planning to design a suite of language-independent evaluation measures for quantifying the success of each dialogue. Establishing reliable automatic evaluation measures for helpdesk dialogues should ultimately enable researchers and engineers to build and tune intelligent helpdesk systems efficiently. This paper reports on our initial step towards this direction to seek feedback from the Japanese natural language processing research community. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
Whenever a user of a commercial product or a service encounters a problem, an effective and efficient way to solve it is to contact the helpdesk. In the present study, we address the problem of evaluating textual dialogues between the customer and the helpdesk, such as those that take the form of online chats. Thus, we consider textual, dyadic and task-oriented dialogues. Successful dialogues are desirable both for the customer and for the company that sells the product / service. We are currently building a small-scale Chinese helpdesk dialogue test collection (with English translations) and planning to design a suite of language-independent evaluation measures for quantifying the success of each dialogue. Establishing reliable automatic evaluation measures for helpdesk dialogues should ultimately enable researchers and engineers to build and tune intelligent helpdesk systems efficiently. This paper reports on our initial step towards this direction to seek feedback from the Japanese natural language processing research community. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AN10115061 |
| 書誌情報 |
研究報告自然言語処理（NL）
巻 2016-NL-228,
号 1,
p. 1-10,
発行日 2016-09-22
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8779 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-NL16228001.pdf (440.6 kB)
