newsN

I M A G E F R O M T H E K E Y W O R D / G O O G L E

ingly good, and it’s all thanks to neural machine translation,” Cho said.

Data posted by Google in November
2016 show that its new system is now
Until a couple of years ago, the
steady progress in machine translation
had always been dominated by Google,
with its well-supported phrase-based
statistical analysis, said Kyunghyun
Cho, an assistant professor of com-
puter science and data science at New
York University (NYU).

However, in 2015, Cho (then a
post-doc in Yoshua Bengio’s group
at the University of Montreal) and
others brought neural-network-
based statistical approaches to
the annual Workshop on Machine
Translation (WMT 15), and for the
first time, the “Google translation
was not doing better than any of
those academic systems.”
Since then, “Google has been re-
ally quick in adapting this (neural
network) technology” for translation,
Cho observed. Based on its success,
last fall Google began replacing the
phrase-based system it had used for
years, starting with some popular lan-
guage pairs. The new system is “amaz-
Deep Learning Takes
on Translation

Improvements in hardware, the availability of massive amounts of data, and algorithmic upgrades are among the factors supporting better machine translation.

Science | DOI: 10.1145/3077229 Don Monroe

NEW TRANSLATION OLD TRANSLATION

A demonstration of the improvement in Google Translate thanks to the use of neural machine translation.