2019 Fiscal Year Final Research Report
Acquiring domain knowledge during dialogues for dialogue systems
| Project/Area Number |
16H02869
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
武田 龍 大阪大学, 産業科学研究所, 准教授 (20749527)
林 克彦 東京大学, 大学院情報理工学系研究科, 助教 (50725794)
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Keywords | 対話システム / ドメイン知識獲得 / 未知語 / 未知概念 / 暗黙的確認 / 質問選択 |
|---|
| Outline of Final Research Achievements |
We have studied towards an intelligent dialogue system that autonomously acquires domain knowledge and becomes smarter as we talk. Specifically, we proposed a process called implicit confirmation for acquiring knowledge about unknown terms that appear during dialogues, without reducing the user's willingness to talk. We also investigated and revealed the influence of the question types to acquire knowledge on the user's impression using crowdsourcing. Furthermore, we proposed a novel framework for selecting questions about knowledge unknown to the system, using a technique called knowledge graph completion.
|
| Free Research Field |
知能情報学
|
| Academic Significance and Societal Importance of the Research Achievements |
対話システムが話しながら知識を獲得することは重要なトピックであり，ここ数年で海外のいくつかのグループでも研究が開始されている．その中でも我々のグループでは，ユーザの話す意欲を削がない手法に注力しており，このような観点からのアプローチは世界的にもまだ類を見ない．一見賢いがすぐに飽きられるような対話システムの次世代を考えるうえで，個人に応じてシステムが育つ技術は必須であり，それを実現するための基盤技術にあたる．
|
