META-NET Looks to Neural Machine Translation

Map of the languages of Europe. (Source: Wikipedia)

META is the Multilingual Europe Technology Alliance, created by META-NET, an international organization founded in 2010, now comprised of 60 research centers from 34 countries. META-NET’s mission is to foster a language technology foundation for Europe.

A recent article published in Slator, How Neural Machine Translation Can Unlock Europe’s Digital Single Market, was written by Georg Rehm, Rico Sennrich, and Jan Hajič of META-NET. In it they share their opinions and observations of the challenges and opportunities facing the EU.

In keeping with the EC’s priority to create a Digital Single Market, the trio from META-NET noted, “If customers are hampered by language, online commerce will remain confined to fragmented markets, defined and restricted by language silos. Even the unacceptable suggestion for everyone to use English would not deliver a single market, since less than 50% of the EU’s population speaks English, and less than 10% of non-native speakers are proficient enough to use English for online commerce.”

The article goes on to emphasize that limiting localization to German, French, Italian and English only address half the EU’s population. The rest speak one (or more) of the other 20 official EU languages. They point out, “Language barriers in the online world can only be overcome by (1) significantly improving one’s own skills in non-native languages, (2) making use of others’ language skills, or (3) through digital technologies. With the 24 official EU languages and dozens of additional languages, relying on the first two options alone is neither realistic nor feasible.”

While most large enterprises have already overcome localization hurdles in the main languages of the EU, the real challenge remains, in META-NET’s opinion, at the realm of the Small-to-Medium Enterprise (SME) level, and in this other half of the linguistic landscape of Europe.

The article asserts relying on traditional human translation alone would be cost-prohibitive. To reach them will require machine translation, far beyond the quality of today’s tools. “The free machine translation services offered by a few tech giants are useful for giving users the gist of web content. But they cannot be easily and cheaply tailored to support the niche communication needs between SMEs and their customers.”

They propose that a new generation of neural machine translation systems will, in time, provide better translation than current phrase-based statistical model systems. The article bears reading in full.

What are your thoughts on the current state and future of machine translation? What linguistic communities in Europe or beyond do you believe need to be better-served? We’d love to hear! Contact us at projects@e2f.com, and let us know.

Previous
Previous

Localization & Semantic Overload

Next
Next

e2f Wants YOU!