[lg policy] Artificial intelligence goes bilingual—without a dictionary

Dave Sayers dave.sayers at cantab.net
Wed Nov 29 04:27:32 EST 2017

Interesting development about two new automatic translation methods, one developed by 
someone at Facebook:

 From the article:
"In the only directly comparable results between the two papers—translating between 
English and French text drawn from the same set of about 30 million sentences—both 
achieved a bilingual evaluation understudy score (used to measure the accuracy of 
translations) of about 15 in both directions. That’s not as high as Google Translate, 
a supervised method that scores about 40, or humans, who can score more than 50, but 
it’s better than word-for-word translation. The authors say the systems could easily 
be improved by becoming semisupervised–having a few thousand parallel sentences added 
to their training."

The two research papers are: https://arxiv.org/abs/1711.00043, 
https://arxiv.org/abs/1710.11041. (Cautionary note: neither has yet been peer-reviewed.)

Facebook is also very interested in 'augmented reality' (wearable technology which 
overlays what you see and hear with additional/replacement information). So this 
feels like part of the trajectory I described back in April, about the future 
development of wearable automated translation:

Interesting times!


Dr. Dave Sayers, ORCID no. 0000-0003-1124-7132
Senior Lecturer, Dept Humanities, Sheffield Hallam University | www.shu.ac.uk
Honorary Research Fellow, Cardiff University & WISERD | www.wiserd.ac.uk
Communications Secretary, BAAL Language Policy group | www.langpol.ac.uk
dave.sayers at cantab.net | http://shu.academia.edu/DaveSayers

This message came to you by way of the lgpolicy-list mailing list
lgpolicy-list at groups.sas.upenn.edu
To manage your subscription unsubscribe, or arrange digest format: https://groups.sas.upenn.edu/mailman/listinfo/lgpolicy-list

More information about the Lgpolicy-list mailing list