A Report On Machine Translation And Post-editing

From time immemorial,language has been the main carrier of cultural exchange,thus translation between different languages has become critical.However,facing a large number of multi-language information,traditional human translation has been far from satisfying the demand.Therefore,machine translation technology has emerged.Machine translation is the process of translating a natural language into another language automatically by computer.It has developed from rule-based machine translation,corpus-based machine translation to current neural machine translation,and Google neural machine translation is the most representative.According to machine translation results of excerpts of a book,named Marketing Communications in Tourism & Hospitality: concepts,strategies and cases,a non-literary material,this report analyzed the error pattern of Google machine translation.The concrete steps are using Google machine translation(Google Translate)to translate the material first and then conducting post-editing by the researcher.On the basis of functional equivalence theory,the results of Google machine translation are contrasted with that of post-editing so as to analyze the error pattern of Google machine translation.And the rest of the book will be translated and analyzed in a similar way by my partner in this project.It is analyzed both in lexical level and syntactic level,even though Google machine translation has been a milestone breakthrough in machine translation history,mistakes still exists in terminology,polysemy,part of speech,word order,noun phrases,verb phrases,etc.The significance of comparison between machine translation and human translation is to find these typical sentence patterns that need to be formalized,to limit the scope of word processing,to improve the efficiency of language processing in machine translation and to reduce unnecessary waste.The researcher also hopes to improve her own translation skills and accumulate experiences of translation.