One day my grandfather told my dad, that in future water will be sold in plastic bottles. In early 60’s Poland my father burst out laughing.
When I first saw “Star Wars: New Hope” about twenty years ago, Iwould never assume that C-3PO-like translation will be possible (well, Iwould also never thought, that Disney will make seventh episode). Nowadays we have so fast technological expansion, that when I heard about technology that allows real-time communication in different languages – I almost ignored it.
In this article you can read about new Skype feature, you’ll learn how it can help deaf people and why is it important to your international customer support. Take your time!
Few weeks ago, Microsoft announced that Skype Translator will roll-out at the end of summer 2015. Skype Translator will be new feature in Skype, which make possible one-on-one communication between people speaking in different languages. It is being developed as collaboration between Skype and Microsoft Translator teams.
Basically, it works by recognizing your speech and converting it into transcript, and then on the other side it’s spoken by synthesizer in chosen language. Everything goes almost in real-time.
You can try it out already. From 20th may 2014 there is Skype Translator Preview version available for download. It allows you to test speaking communication in six languages: english, french, german, italian, chinese and spanish. Written message translation is available in almost 50 languages.
How is it possible?
We must agree, that it’s a big thing to achieve something like this. So how it was possible?
For few years now we have near-perfect working Google translation of written text. Spoken language is slightly different, as translator need to overcome all the “Uh’s”, “Om’s” and “You-know’s” which are placeholders or even separators in everyday conversations.
First approach was to simply map text translations with their spoken equivalents by connecting between languages not only phrases, but also single words. Data was fed to machine-learning system, called statistical machine translation (SMT) and algorithm dealt with matching phrases.
There’s problem when words in phrase are reordered. SMT doesn’t understand grammar, so it can’t shift order between languages. For example, an english sentence is usually: subject+verb+object, but hawaiian sentence would be like: verb+subject+object.
This is why Microsoft developers went a little further and pioneered syntactically informed SMT. When words are reordered it can break up sentence into single words, match them with their equivalents and connect individual words in correct grammatic order.
Cutting up phrases and connecting individual words may sound like a primitive approach, but it’s not. “That’s pretty much the best method,” says Chris Manning, professor of linguistics and computer science at Stanford. “Microsoft’s machine translation team has been one of the prominent developers in this area, and basically, that is the state of the art in machine translation at the moment.”
Syntactic SMT was huge achievement, but still it could use some improvements. Lots of written-text pages previously fed to system, cannot help with understanding talk, emotions and information between words. This straightly led to analysis of social media conversations. Facebook, Twitter and SMS messages was analysed to find those real-life communication patterns. And they found it.
It’s amazing how fast technologies are progressing. The ability to speak with everyone around the globe would be most life-changing feature since popularisation of smartphones. Imagine that you don’t need to know foreign language to go to any country in the world, because you’ll be able to communicate with anyone using your mobile device.
But let me give you example of different approach which is even more touching.
One of Microsoft researchers, Ted Hart, happens to be deaf. He lost his hearing, when he was thirteen years old. When he met his wife, they communicated only by drawing letters on his hand. After some time, his wife learned the sign language to enter Ted’s world. He never had opportunity to call his wife phone number.
Skype Translator changed this. As it is transcripting speech, Ted had opportunity to call his wife and see almost in real-time what she’s saying. Watch video about that story!
Great, isn’t it? And that’s not the end of unlimited possibilities it can bring. Imagine world without all language barriers!
Speaking about world without language barriers we’re coming to the point about your service/product. How do you provide support for your customers? For now you probably reply to tons of emails, make a phone calls, communicate trough communicators (also Skype) even schedule face-to-face appointments.
But still – it’s limited to only those languages which your employees know.
Providing support in new languages require to hire employee with that language, or send one of them to language course (which can take years to learn!). That costs, and cost is something what you want to avoid. Look again at Skype translator. It will allow you to provide support or sale all over the world. Well, that sounds right.
To sum up, Skype Translator is still self-learning on the data which it gathers and at this moment it’s still in development. I believe that it’ll revolutionize the way of our communication. If you are using Windows 8.1 or Windows 10 you can try it out yourself today.
So here’s the question: would you trust it enough to communicate with your possible partners? Are you willing to use it to expand into new markets?
I’d love to hear your thoughts!
[do action=”cc-image-attribution” author=”jdhancock” photourl=”https://www.flickr.com/photos/jdhancock/5703091404/” cclicense=”by”/]
[do action=”cc-image-attribution” author=”wiertz” photourl=”https://www.flickr.com/photos/wiertz/13195153284″ cclicense=”by”/]
[do action=”cc-image-attribution” author=”mstable” photourl=”https://www.flickr.com/photos/mstable/17493717036″ cclicense=”by”/]
[do action=”cc-image-attribution” author=”jlhopgood” photourl=”https://www.flickr.com/photos/jlhopgood/6653628559″ cclicense=”by-nd”/]