Problems And Recommendations For Ai Application In E-C Consecutive Interpreting In Informal Language

Artificial intelligence(AI)has become an increasingly hot topic in recent years.Machine translation as an important branch of natural language processing,is also one of the ultimate targets of AI.In September 2016,Google introduced the first neural machine translation system called GNMT.Against this background,this thesis aims to study the problems of and put forward recommendations for AI application in E-C consecutive interpreting in the context of informal language.This thesis chooses Google Translate as the research object for interpreting three different sources of English audio materials featuring informal spoken language and keeps records of the interpreting transcripts of the app.The research method applied in this study is text analysis based on source language(excerpts from material transcripts)and target language(interpreting contributed by Google Translate).Altogether seventeen examples are extracted for the analyses.By thoroughly going over each example,the author finds GNMT is hardly able to interpret English into idiomatic,sense-consistent and fluent Chinese,especially with long sentences and paragraphs.The author therefore classifies the problems with AI interpreting in informal language from English to Chinese into four categories,i.e.,inaccuracy of terms and words,syntax error,lack of perception in context and lack of fluency in delivery.As the majority of problems come with grammar and syntax mistakes,the recommendations raised thus are devoted more to this respect.These recommendations are primary ones underpinned by linguistic knowledge.They include,high quality big data input of informal language,interpreting localization and out-of-domain adaptation.This thesis is divided into five parts as follows:introduction,literature review,case description,case analysis and conclusion.Chapter one(Introduction)has a brief introduction to the background of AI application in interpreting and the purpose and value of this thesis.Chapter two(Literature Review)goes over fundamental views,development and previous studies of AI interpreting as well as studies on informal language interpreting.Chapter three(Research Design)gives a brief profile of the application Google Translate and describes the research design.Chapter four(Case Analyses)points out the problems of AI application in E-C consecutive interpreting of informal language and presents recommendations by analyzing some examples interpreted by Google Translate.Chapter five(Conclusion)summarizes the major findings of this thesis and illuminates the merits as well as limitations.The problems found through text analysis can be regarded as preliminary reference for further studies.