Publications

Publication details [#47122]

Zhao, Shengfang. 2021. Post-editing Neural Machine Translation Versus Human Translation for Chinese Essays: a pilot study. In Moratto, Riccardo and Martin Woesler, eds. Diverse Voices in Chinese Translation and Interpreting: theory and practice (New Frontiers in Translation Studies). Cham: Springer. pp. 399–432.

Abstract

With the rise of neural machine translation (NMT) and encouragement of previous research, a pilot study has been conducted to compare post-editing (PE) Google neural machine translation (GNMT) to human translation (HT) with Selected Modern Chinese Essays: Annotated Bilingual Edition, Volume 1. To this end, six student translators were invited to take part in a voluntary experiment. They were in M.A. Chinese-English translation postgraduate program with three of them having up to two years’ professional translation experience. They were asked to translate two Chinese essays in two workflows: post-editing Google neural machine translation and translating from scratch, respectively. After the experiment, the translation time taken was analyzed; the qualities of the translations were compared by the author. Results indicate that PE was faster than translation from scratch for all participants but one. The post-edited translations were found to be of inferior quality compared to human translations produced from scratch, but this should modulate the productivity benefits of PE quite considerably.
Source : Based on publisher information