Machine Translation Quality Assessment:Methods,Applications,and Outlook
The widespread adoption of machine translation post-editing underscores the importance of translation quality evaluation.This paper examines three primary evaluation methods(human evaluation,reference-based automatic evaluation,reference-free automatic estimation),delineating their strengths and weaknesses.It advocates adaptable selection of evaluation methods according to specific requirements.Future research should prioritize in enhancing the accuracy of automatic evaluation,broadening evaluation dimensions,and conducting domain-specific evaluations.
NETL
NSTL
万方数据
