Our Analysts' Insights

The explosion of enthusiasm for MT that followed Google’s announcements about neural machine translation (NMT) last year shows no signs of abating.

In the last year, CSA Research has been covering a new paradigm for translator productivity, one we call “augmented translation.” The term comes from “augmented reality”: applications that overlay images of the world with relevant information.

Imagine yourself in a café in Paris or on a beach in Cancún, running into some gorgeous human specimen you just can’t help but approach. You walk up to the person, offer them a hearing device, and point for them to put it in their ear while you pop one in yourself. Then you launch an app on your smartphone and start communicating via the help of machine interpreting, hoping the app will accurately translate your best pick-up line.

Every new development in the field of machine translation (MT) is accompanied by a set of numbers that purport to show dramatic improvement in terms of quality, usually BLEU or METEOR scores. These measures use a scale from 0 to 100 to quantify how similar the MT output is to one or more human translations of the same source text based on a mechanical analysis of how many of the same words show up and how likely they are to appear in the same order.

Language services today stand on the cusp of a disruptive transformation that will redefine how professional linguists work. This shift will come from the availability of ubiquitous artificial intelligence (AI) that extends their reach and capability and makes them far more efficient than they could otherwise be.

Recent advances in neural machine translation (NMT) represent a significant step forward in machine translation capabilities. Although most media coverage has significantly oversold the technology, one of Google’s announcements may actually be the most important one in the long run – the first successful deployment of zero-shot translation (ZST).

By now, most language professionals have seen that claims that neural machine translation (NMT) is delivering results as good – or almost as good – as human translation. If these claims – which have been repeated in the mainstream tech press without much examination – are accurate, it is only a matter of time before human translators will be out of work.

SYSTRAN announced the beta test of its Pure Neural Machine Translation (PNMT) software with 30 language pairs (18 with English, 12 with French), a dozen corporate clients in diverse industries, and online public access to its software. This beta program caps a year of industry and media attention to deep learning, artificial intelligence, and more recently neural MT (NMT). SYSTRAN announced its PNMT product in August with this October beta. Google revealed its single-pair NMT solution in late Se...

Machine translation (MT) has caught the public eye once again. The Wall Street Journal recently predicted that “the language barrier is about to fall” – within 10 years. We might note that this is one barrier that has been “about to fall” for far longer than the actual Berlin Wall stayed up, and that it’s been an awfully long 10 years since the first such claims were made in the 1950s.

Better machine translation (MT) is surely at the top of many organizations’ globalization wishlists. But they all wonder how they can develop good MT engines that produce usable results.