Jul 11, 2018

In late June SDL , a UK-based developer of language translation technologies, announced that its solution, a next-gen Neural Machine Translation 2.0, has learned to translate from Russian to English with “over 90% of its output labelled as perfect by professional Russian-English translators.” “A machine running SDL NMT 2.0 can now produce translations of Russian text virtually indistinguishable from what Russian-English bilingual humans can produce,” said Samad Echihabi, Head of Machine Learning R&D at SDL. Currently the solution is available via a special server . The linguistic properties and intricacies of the Russian language relative to English make it particularly challenging for MT systems to model. Since Russian is a highly inflected language with different syntax, grammar, and word order compared to English, raising the translation quality has been an ongoing focus of the SDL Machine Learning R&D team, says SDL. “One of the toughest linguistic challenges facing the machine translation community has been overcome by our team,” said Adolfo Hernandez, CEO at SDL, adding that now the company “is paving the way for leadership in other complex languages.” Created more than 25 years ago, LSE-listed SDL presents itself as a global leader in innovative language translation technology, services and content management. 1