A Report On Machine Translation + Post-Editing Of The History Of National Southwest Associated University

With the development of artificial intelligence and the growing demand for translation services,machine translation(MT)has become widely adopted.However,due to the complexity of translation practice,MT output often falls short in terms of quality compared to human translation.As a result,the machine translation +post-editing(MTPE)model has emerged and is now widely used in the language service industry.In this report,the MTPE translation model is integrated with the widely adopted Dynamic Quality Framework and Multidimensional Quality Metrics(DQF-MQM)for assessing translation quality in the language service industry.To ensure high-quality translations,the author evaluates MT output,identifies the types of errors and proposes post-editing strategies.As for the source text(ST),the initial two sections of Part Three have been selected from The History of National Southwest Associated University: Peking,Tsinghua and Nankai from 1937 to 1946.This is an informative text contains numerous high-frequency terms and is well-suited for processing by MTPE model.The author employs Youdao Online Translation,developed by Net Ease,to generate MT output.The quality of the MT output is evaluated according to the DQF-MQM framework before analyzing and summarizing the main types of errors.Based on specific case analysis,countermeasures for post-editing are proposed.This research aims to provide a reference for applying the MTPE model in the language service industry.According to the DQF-MQM framework,errors in this MT output can be mainly categorized into two types: errors of accuracy and errors of fluency.The report focuses on mistranslations of proper nouns and sentences for accuracy,while inconsistency and grammar are the main focuses for fluency.Through case analysis,this paper further explores corresponding post-editing strategies.