2001 Fiscal Year Final Research Report Summary
Super-Function Based Japanese-Chinese-English Machine Translation and Information Retrieval
| Project/Area Number |
11480080
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | The University of Tokushima (2001)
Hiroshima City University (1999-2000) |
|---|
Principal Investigator |
FUJII Ren The University of Tokushima, Faculty of Engineering, Professor, 工学部, 教授 (20264947)
|
|---|
| Project Period (FY) |
1999 – 2001
|
| Keywords | Machine Translation / Information Retrieval / Muchi-Language Processing / Natural Language Processing / Super Function / Chinese / Japanese / English |
|---|
| Research Abstract |
The aim of this project is that studying a new paradigm of Machine Translation and Information Retrieval to address most users' requests. Over last years, we have paid attention to the favors of MT users and found that most users prefer that the system is fast, inexpensive, easy to control, and easy to update. In this project, we proposed a new MT paradigm called Super-Function based Machine Translation (SFMT). SFMT uses Super-Function to translate without thorough syntactic and semantic analysis as most MT systems usually do. This year, we have built a prototype system based on Super-Function. Some Super-Functions both Japanese-English and Japanese-Chinese have been acquired from parallel corpora, and some experiments on MT and IR have been carried out. The following shows the result of our research.
1. Some Japanese-English super-functions have been extracted and the generation method was presented.
2. Research on Automatic Summarization and Multiple Engines have been done.
3. Prototype English-Japanese machine translation system based super-function has been built.
4. A translation engine using super-function for Web user is developed,
5. One algorithm of IR using super-function is proposed.
6. A new concept of Multi-Language Multi-Function Multi-Media intelligence system is presented and a prototype IR system using super-function has been constructed. The result of our experiments show the proposed method is efficient.
|
