Books & Rights Marketplace 
日本語  |  ENGLISH 
Welcome Question?
Privacy Policy | Terms & Conditions  
First time user?
Please register online today !  
Register Now! 
MemberLog-in: Password: Login
 
 
Affiliated Partners
 
SEARCH
Topic
Author(Surname by initial)
 
ACCESS RANKING
This week's Top 10 Ranking 
1.  LOOPS ― The Seven ...
2.  The spiritual way ...
 

Become a fan of BRM


Promote Your Page Too
 
 
News & Events
2018.05.15 Quality of Chinese-to-English AI Translation Reaches Human Parity
According to the featured piece on Microsoft's "The AI Blog" dated March 14, 2018, the quality of Chinese-to-English machine-translated news articles has reached “human parity.” To check quality and accuracy, bilingual evaluators compared human and AI (artificial intelligence) translations, and it was concluded that both were at the same level. 

A test set called “newstest2017”, released at a machine translation conference called WMT17 in 2017, was used in this project which involved researchers in Beijing and Washington. It extracts about 2000 sentences from a collected sample of online news articles to use for accuracy testing in machine translation.

Deep learning was adopted to train the AI. To increase the translation accuracy, researchers trained the AI through various methods, including those that resemble human learning techniques. For example, dual learning, where the system re-translates the text back to Chinese from English, and deliberation networks, where the system repeatedly translates the same sentence, were used. These training methods are also expected to play a role in furthering AI technology—in translating language pairs besides Chinese and English, in translation other than news articles, and finally, in fields outside of translation.

The project’s success is a groundbreaking development in machine translation. It may be the start of using AI translation to quickly and accurately disseminate information around the world. Of course, despite the many years of research and development going into improving the accuracy of machine translation, many issues remain. It's said that the next goals will be increasing the number of language pairs, real-time news translation, and human-level speech-to-speech translation. 

For details, please visit:
 
http://techwireasia.com/2018/03/microsoft-developed-bot-can-apparently-translate-well-human/

https://blogs.microsoft.com/ai/machine-translation-news-test-set-human-parity/

https://japan.cnet.com/article/35116178/