A Study On The Quality Of C-E Machine. Interpreting

As translation technology advances,there has been considerable controversy about whether machine interpreting can ever fully replace human interpreting and about the quality of machine interpreting.Faced with such a predicament,it seems that there would be a need for doing research on the quality of machine interpreting in order to know the gap between human and machine interpreting.In literature review,the majority of research in this field is conducted from the perspective of computational science or computational linguistics,and there is not much research on the quality of machine interpreting based on its own quality criteria,and from the perspective of particular linguistic features.Additionally,the analysis of machine interpreting is largely undertaken with the quality criteria of human interpreting,and the study of machine interpreting with its own criteria is just the story in recent years.Based on this recognition,the author chooses “completeness,expression and grammar” as the quality criteria after a review of previous studies on the development of machine interpreting and quality criteria of interpreting.This thesis takes Zhang Lu’s interpreting in the conference of Premier Li Keqiang meeting the press after the Two Sessions in 2021 as the analytical reference and conducts a consecutive interpreting using i FLYTEK and Sougou Translate,taking Zhang Lu’s version to conduct a quality analysis of machine interpreting.This study finds that low accuracy of speech recognition,overliteral translation on four-character phrases and metaphorical expression can violate the criterion of“completeness”.Poor performance can also be found in information delivery which fails meet the requirement of the criteria “expression”.In terms of “grammar”,machine interpreting is unable to use correct form of singular and plural,additionally,machine cannot distinguish the subject clearly.