We tried translating the press release written for the company (original text is Japanese) by using machine translation + post editing.
According to the staff who was in charge of the MT + PE,
"I believe that the accuracy of the MT was about 7 out 10. I felt that company B had more natural translation than company A. When translating term by term, the suggested synonyms were more appropriate (company B more than company A).
Further, the staff thinks that "the productivity increased about 30 to 35% overall".
With patent text, the productivity increased by 20 to 25%, so it seems that productivity increases more with business texts.
If we can increase this much translation efficiency in total, this may amount to a daily worth of workload per staff.
Currently, we are using machine translation engine from two companies. We will surely continue the verifying process.
Above is translation of an article "１０段階で７でした。" dated October 6, 2017
Translation by Hiroko Matsuda
With patent text, the productivity increased by 20 to 25%, so it seems that productivity increases more with business texts.
If we can increase this much translation efficiency in total, this may amount to a daily worth of workload per staff.
Currently, we are using machine translation engine from two companies. We will surely continue the verifying process.
0 件のコメント:
コメントを投稿