A Quantitative Study And Comparison Of Foreign Affairs Translation System,google And Baidu Machine Translation Systems In Translating Foreign Affairs Texts

In 2016,there are numerous reports of breakthrough in machine translation technology from home and abroad.Whether machine translation can replace human translation has become one of the hit topics in academic community.In this thesis,it advocates that in some specific translation cases,machine translation systems can produce high quality translation texts and improve efficiency of translation activities.But in general,machine translation results are still unsatisfied.It is impossible to replace human translation within a short time.The author has collected 100 articles from released foreign affairs texts and their relative translation of Ministry of Foreign Affairs of the People's Republic of China as research materials.Then the author has compared three representative machine translation results with the standardized human translation.It is aimed to find out differences of machine translation systems in lexical density,mean sentences through quantitative analysis and case studies.In September 2016,Google released Neural Machine Translation System which has improved the quality of machine translation.And in November Robin Li,CEO of Baidu,said on the 3rd World Internet Conference Wuzhen Summit that the automatic translation from Baidu Artificial Intelligence would be put into practice.Therefore,the author can get machine translation results from both home and abroad representative machine translation systems specifically “Google” and “Baidu” with general corpus.During the internship in the Information Center of Foreign Affairs Office of the People's Government of Beijing Municipality,the author also got machine translation versions by its inner Foreign Affairs Translation System with specialized corpus.Then,the author has built a comparable corpus between a human translation and three machine translation results of thousands of sentences.By the use of statistical tools like WordSmith,CLAWS and detailed examples,the author has summarized differences on terms,synonymous,grammar and signs translation between machine translation systems,which will provide corpus proof to improve the quality of machine translation.This thesis aims to compare and analyze advantages and disadvantages of the Foreign Affairs Translation System,Google and Baidu machine translation systems.It has also provided solutions to address problems and made contributions for future development of machine translation systems.