2012 Fiscal Year Final Research Report

A Study on Translation Repair Method Optimized for Use Situation

Research Project

Project/Area Number	23800014
Research Category	Grant-in-Aid for Research Activity Start-up
Allocation Type	Single-year Grants
Research Field	Media informatics/Database
Research Institution	The University of Tokyo
Principal Investigator	MIYABE Mai 東京大学, 知の構造化センター, 特任研究員 (00613499)
Project Period (FY)	2011 – 2012
Keywords	多言語間コミュニケーション / 機械翻訳 / 多言語処理
Research Abstract	The purpose of this research is to develop translation repair method optimized for use situation, and improve the reliability of multilingual communication via machine translation. The contribution of this research is as follows: (1) this research revealed that machine-translated sentences are available in the medical field if they have information of their original sentences without exaggeration and without omission. (2) We developed a prototype system using detection methods of breaking the rules for accuracy translation. Moreover, we proposed a method for detecting the mismatch between a target-translated sentence and its back-translated sentence using multiple translation systems. This method can detect approximately 71% of the mismatches.

(3 results)