Automatic translation of written text has become fairly ubiquitous and commonplace on the internet- you’d be hard pushed to find someone who hasn’t leaned on Google or Bing translators once in a while. But now deep learning technology means that instant translations of speech may come to be just as commonplace. Microsoft have been working on real-time speech translation using deep-learning technologies for a while. The result? Skype Translate, which translates your speech as you talk, and be available as a Windows 8 beta App at the end of the year.
Microsoft CEO Satya Nadella seemed proud but somewhat mystified by just how intelligent their deep-learning system had become. “You teach in English, it learns English. Then you teach it Mandarin — it learns Mandarin, but it becomes better at English. And then you teach it Spanish and it gets good at Spanish, but it gets great at both Mandarin and English. And quite frankly none of us know exactly why.”
Nadella and Gurdeep Pall, the head of Microsoft’s Skype and Lync division, demo-ed Skype Translate at Code Conference 2014 on Tuesday. English- speaking Pall had a conversation with a German colleague; once one of them had spoken, Skype would translate and ‘read out’ the dialogue in the other’s native tongue. Although not 100% accurate, the translation was precise and clear enough to be seriously useful.
The same technology already underpins the Bing Translator and the speech recognition capabilities of the Cortana personal assistant in Windows Phone 8.1. Which so much material to learn from, the deep-learning technology could turn Skype Translate into an incredibly accurate and powerful tool.
Read more here.
(Photo credit: Patrick Haney)
