A Comparative Quality Assessment Of Human Interpretation And Machine Interpretation Based On The Quantitative And Qualitative Assessment Theory

Google's Neural Machine Translation(GNMT)system,developed in September 2016 by the US technology company Google,is said to have greatly reduced English-Chinese and Chinese-English translation errors compared with the traditional Phrase-Based Machine Translation(PBMT)system.This news has generated a lot of public discussions on whether or not machines will fully replace human translators and interpreters.Since the GNMT system came out not long ago,few researches have been conducted on GNMT's translation and interpretation quality assessment.Moreover,existing researches mainly focus on theoretical standards and models,limited by a lack of quantitative analysis and complete methodlogies.This thesis,however,based on Yang Chengshu's Quantitative and Qualitative Assessment Theory,aims to take into account both reliability and validity,and to make a comparative analysis of human interpretation and machine interpretation through quantitative evaluation and qualitative description in a scientific and fair manner.This thesis studies four groups of questions and answers between foreign journalists and Chinese Premier Li Keqiang at the 2017 press conference,compares the senior interpreter Ms.Zhang Lu's English-Chinese and Chinese-English consecutive interpretation quality with that of the mobile application Google Translate from the perspectives of fidelity,delivery,language,and time control by using specific quantitative evaluation and qualitative description indicators,and then summarizes their respective strengths and weaknesses.This thesis finds that Ms.Zhang Lu's human interpretation enjoys a better overall quality than the machine interpretation of Google Translate,with higher fidelity in content,higher expressiveness in delivery,higher accuracy in language,and higher flexibility in processing information,but with more omissions,more over interpretations,and lower fluency.This research suggests that interpreting students and teachers and professional interpreters should improve both strengths and weaknesses,explore more possibilities of machine-aided human interpretation,and take a positive attitude towards opportunities and challenges brought by machines and artificial intelligence to translation and interpretation talent training and the language service industry.