專家聲稱 5到10年內機器翻譯全面普及

Experts believe that machine translation good enough to replace advanced human translators will be widespread within five to 10 years, as language-processing technology has seen many breakthroughs in recent years, reported on May 22.

據新華網5月22日報道，由於近年來語言處理技術取得了多項突破，專家們相信，足以取代資深譯員的機器翻譯將在五至十年內普及。

Zhou Ming, standing deputy dean of Microsoft Research Asia, made the statement in his keynote speech at the 2017 Global Artificial Intelligence Technology Conference, held in Beijing on May 22.

5月22日，微軟亞洲研究院常務副院長周明在於北京舉行的2017全球人工智能技術大會上發表的主旨演講中這樣說道。

Machine translation currently faces challenges that prevent it from reaching the standards of professional translation, Zhou noted.

周明指出，當前，機器翻譯面臨着各種挑戰，這使得其無法達到專業翻譯的標準。

However, he also pointed out that several technology companies are now focusing on machine translation, and have already made great progress.

不過他也指出，現在好幾家技術公司正在主攻機器翻譯，且已取得很大進展。

Google released its translation system in September 2016, based on recurrent neural networks.

2016年9月，谷歌公司發佈了自己基於遞歸神經網絡的翻譯系統。

In addition, the artificial intelligence team at Facebook claimed to have achieved the highest accuracy yet in machine translation by using brand new neural network technology, with a translating speed nine times that of Google.

此外，臉書的人工智能團隊宣稱，通過使用新型神經網絡技術，已經實現了迄今最爲準確的機器翻譯，翻譯速度爲谷歌的九倍。

Apart from translation, Zhou speculated that language-processing technology, integrated with other artificial intelligence, could be applied in many fields including education, medicine and driverless cars in the near future.

周明預測，除了翻譯，語言處理技術和其它人工智能，在未來還可應用於包括教育、藥物和無人駕駛在內的很多領域。