基于机器翻译的人机交互译前编辑模式研究  

An Exploration into the Pre-editing Model Based on Machine Translation in the Interactive Human-machine Context

作　　者：李洋[1] 孙宁 梁玉静 LI Yang;SUN Ning;LIANG Yujing

机构地区：[1]东北大学外国语学院,辽宁沈阳

出　　处：《语言教育》2021年第2期55-62,共8页Language Education

基　　金：全国翻译专业学位研究生教育研究项目“语块化口译教学方法的探索与实践”(项目编号:MTIJZW202006)的阶段性成果。

摘　　要：译前编辑是提高机器翻译质量的重要途径之一,也是人机交互翻译的重要表现。本研究从《欧盟GMP/GDP法规汇编》中选取案例作为分析对象,通过对比后选用谷歌翻译平台,对未经和经过译前编辑的机译文本进行对比分析,提出在词级和句级层面的译前编辑方法,以期提高人机交互译前编辑模式下机译译文质量。Pre-editing is regarded as one of the important approaches to improve the quality of machine translation,and it is also an important manifestation of the interactive human-machine translation.The current study selected source texts as cases from the Compilation ofE U GMP/GDP Regulations.In what follows,a comparative analysis was scored on the quality of target texts output by different machine translations.As a result,Google Translate outperformed the others.Furthermore,based on both unedited and pre-edited texts that were rendered by Google Translate,the study proposed some pre-edited techniques at lexical and syntactic levels in order to improve the quality of machine translated texts in the interactive human-machine context.

关 键 词：机器翻译 译前编辑模式 译前编辑方法 人机交互 《欧盟GMP/GDP法规汇编》 

分 类 号：H059[语言文字—语言学]