Construction of training system for novice Japanese conversation beased on the communicative approach
| Project/Area Number |
15500618
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Educational technology
|
|---|
| Research Institution | Shizuoka University |
|---|
Principal Investigator |
KONISHI Tatsuhiro Shizuoka University, Faculty of Informatics, Associate Professor, 情報学部, 助教授 (30234800)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
ITOH Yukihiro Shizuoka University, Graduate school of science and technology, Professor, 創造科学技術大学院, 教授 (20193526)
KONDO Mokoto Shizuoka University, Faculty of Informatics, Professor, 情報学部, 教授 (30225627)
|
|---|
| Project Period (FY) |
2003 – 2006
|
| Project Status |
Completed (Fiscal Year 2006)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2006: ¥600,000 (Direct Cost: ¥600,000)
Fiscal Year 2005: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 2004: ¥1,000,000 (Direct Cost: ¥1,000,000)
Fiscal Year 2003: ¥1,200,000 (Direct Cost: ¥1,200,000)
|
|---|
| Keywords | Educational Technology / Computer systems / Information systems / Artificial Intelligence / User interface / Natural language processing / Dialogue systems / Communicative approach / ユーザインタフェース / 日本語教育 / Focus on Form / 誤り検出 / 教育戦略 / 対話訓練システム / 語学教育システム / 自然言語対話システム / 学習者モデル / 知的教育システム / 文脈処理 |
|---|
| Research Abstract |
1. Development of natural language understanding unit
We set the task of searching a hotel as our target task, and performed case study on necessary concepts, vocabularies, and styles of input sentences. According to the result, we design knowledge representation and developed a natural language understanding unit.
2. Development of a method for understanding sentences including wrong expression
We developed an intelligent unit classifying wrong sentences by whether the sentence includes any syntactic mistakes or not, and by whether the sentence can be matched to the context of conversation. Based on this classification, our system can generate adoptive message for each learners.
3. Design and implementation of teaching strategies
We designed a set of teaching strategies for training of novice Japanese dialogue. If a learner input some wrong sentences, the system should consider whether it should point out the learner's mistakes soon, or it should continue dialogue. We adopt a pedagogical method called "Focus on Form" and develop a unit of detecting a certain pattern of words called "Form" from input sentences, in order to realize conversation based on "Focus on Form".
4. Development of an authoring tool
We developed an authoring tool that generates "Situation knowledge" which represents the situation of current conversation, based on natural language sentences representing the situation.
5. Construction of a training system and evaluation of it
We develop a training system for novice Japanese conversation and evaluate the system by testing how well it can detect mistakes included in novice Japanese learners' typical wrong sentences.
6. Publication
We publish a technical report on the result of this research.
|
Report
(5 results)
Research Products
(28 results)
