To Capt. James T. Kirk, and the rest of the crew on Star Trek’s Starship Enterprise, space was the final frontier. To the developers at the Microsoft-owned Skype online videophone service – who are taking inspiration from the hit show’s “space age” technology – real-time translation appears to be the next big challenge.
Microsoft released this week a preview of Skype Translator, which allows video conversations to be translated between English and Spanish in real-time. The full version, which will allow for other languages to be translated, will be available soon.
The idea of an instant universal translator has been a staple of science fiction for decades in many different forms, from the original "Star Trek" television show – on which it was never made explicit where the translator was or how it was constructed – to the "Babel Fish" from Douglas Adams' "The Hitchhiker’s Guide to the Galaxy," which was an actual fish that got inserted into a person's ear.
In contrast the Skype Translator is a computer app that's now available for free download for Windows 8.1. It can translate spoken English and Spanish and more than 40 languages should you choose to instant message instead. Users can sign up for the service, but only if they have latest version of Windows.
"Skype Translator relies on machine learning, which means that the more the technology is used, the smarter it gets," Gurdeep Pall, vice-president of Skype and Lync (Microsoft's IM service), said in a blog post. "We are starting with English and Spanish, and as more people use the Skype Translator preview with these languages, the quality will continually improve."
A video posted on YouTube by Skype shows two middle-school age girls – one in Tacoma, Washington, the other in Mexico City – have a conversation translated over video chat.
The translation service is part of Microsoft’s artificial intelligence research, which is similar to both Google and Apple’s voice activated services in that it relies heavily on machine learning and deep neural networks. The tech giant is also hoping that its translation system can break down language barriers and make communication between cultures easier.
Much like Google Translate, however, the system is not perfect and relies on great amounts of data to form phrases that appear natural. That’s why Skype developers say that the more people that use the preview, the more data will be collected and the better the translations will be.
Each call is broken down into tiny extracts, anonymized and stored on Microsoft’s servers that are then fed into Microsoft’s statistical models to help improve word matching capabilities.
"Skype Translator records conversations in order to analyze the scripts and train the system to better learn each language," Mo Ladha and Chris Wendt, two of Microsoft’s research managers, wrote in a blog post, according to the Guardian. "Participants are all clearly notified as the call begins that their conversation will be recorded and used to improve the quality of Microsoft’s translation and voice recognition services."