Mobile fist, cloud first Microsoft has announced general availability of its Translator for Android wear, Apple watch, as well as Windows Phone, and Windows devices. It supports software including Office 365, Skype, Internet Explorer/Edge, Safari, Bing and Cortana, PDF and offers web developers a translator widget to put on their web sites.
Major software and online shopping providers including Adobe, Amazon, eBay, Etsy, Dell, HP, Facebook, Twitter, Yelp, PayPal and others will include this as part of their customer interface.
Translations are fast. They are either done in the cloud – where latency is more dependent on internet access and speed – or one device via translation packs for off-line translation.
If your device has a camera you can photograph the text. Or you can simply speak directly into the device or type or cut and past the words.
Conversation – as in Skype was the hardest
Skype translator does voice and typed messages. I have seen this in real time and it works well. It was released in May this year and represents decades of work by researchers to deliver real-time speech translation to life.
“Conversational training data is hard to obtain, and the team had to develop new techniques to collect conversational speech data. Social-media posts gives us exposure to yet another genre. How people write on social media is not how they speak, but still, there’s some crossover of slang and related utterances that can help this system and make it contemporary,” says Microsoft’s Arul Menezes.
There’s also the issue of “disfluency,” the difference between how people write and talk. When talking, people use lots of pauses and Ummm - meaningless utterances that bridge gaps between their thoughts. “It’s not just repeating a single word,” Menezes says. “Sometimes, you’ll go three words into a sentence and then back up and restate it. In some languages, it’s more of a challenge than others, especially languages like Spanish, where words have to agree in grammatical gender.”
Untangling such conversations requires lots of training. So does determining where sentence breaks occur. The sentence is the basic unit in translation, and without punctuation, it can be difficult to identify. In fact, it can be next to impossible to read a transcript of a conversation lacking punctuation. The translator must learn to segment out the speech input, too.
“That’s one of the things over the last year that my team’s been doing, resolving the mismatch between the way people talk and the way they write,” Menezes says. “If your translation system is focused on written text, it works very poorly with spoken language.”
Developers can use it
Microsoft Translator API is a cloud-based automatic translation (a.k.a. machine translation) service. It powers translation features across Microsoft products, including Office, SharePoint, Lync, Yammer, Visual Studio, Bing, and Skype. Simply integrate translation into web, desktop, or mobile applications. The API provides a rich functionality set and a wide choice of interfaces: AJAX, HTTP, SOAP, OData, and the Translator Web Widget.