New brain under old hat [DNA : Sunday]
(DNA : Sunday Via Acquire Media NewsEdge) "Like a beat-up jalopy, gadgets that could recognise and respond to voice commands have bobbed around tech streets for quite some time, without anybody ever noticing. Reason: Although VRT or voice recognition technology has, for years, pumped up the pulse rates of techies, in real life it's been a piece of dud. But with the latest version of Google's Android mobile operating system that's all set to change.
And why do we say that The answer lies in an all new VRT system that Google has implanted into the heart of its latest Droid OS, based on what's called a neural network, a computerised learning system that mimics the human brain. And if the breathless excitement of Vincent Vanhoucke, the Google geek who drove the effort, is any indication, then it is quite possible that the old hat does in fact come out looking pret atop its new Droid brain. Says he: "It kind of came as a surprise that we could do so much better by just changing the model."
For starters, voice error rates, the reason why most people never bother with voice commands, are down by 25% in the neural OS, also known as Jelly Bean, according to the latest test reports. This is an important advancement because, modern day gadget users 'speak' to a gadget the way they would to a 'human' and not a robot-using natural language.
Says Arvind Satam, an aural technology evangelist and lead researcher at Microsoft: "The learning curve for gadgets is over. Now, they go extinct if they are unable to understand and behave the way humans do in real time." Be that as it may, the arrival of neural technologies is inspiring a comeback of sorts for VRT, a technology that most tech companies had consigned to the trash can after a decade of hope and despair. In its latest edition the Wired magazine says: "This field of study had cooled for many years, after spending the 1980s as one of the hottest areas of research, but now it's back, with Microsoft and IBM joining Google in exploring some very real applications."
The new Android OS sends the spectrogram of your speech to eight different computers housed in Google's worldwide army of servers-all of them built around Vanhoucke's neural network. These intelligent machines make sense of your voice, much like a brain would, and transmit the information back to your gadgets at lightning fast speeds to ensure you get the response you are looking for. "It's now really like talking with a friend in the same language," says Arvind. "Neural networks are the most exciting thing to happen in technology in some time because like the brain they are programmed to analyze huge number of voice patterns and then predict what a brand new pattern (speech) might mean." In short, like Wired points out,Android takes a picture of the voice command and Google processes it using its neural network model to figure out what's being said.
Neural network algorithms can be used to analyze images too. "What you want to do is find little pieces of structure in the pixels, like for example like an edge in the image," says Hinton. "You might have a layer of feature-detectors that detect things like little edges. And then once you've done that you have another layer of feature detectors that detect little combinations of edges like until the system figures out the entire image," says Vanhoucke.
So, what does all this mean to your life and say the way you use your phone or tab "A heck of a lot," says Arvind. "It could completely turn around the concept of using a gadget," he says. The most obvious change would be that whether you want to send a message or a picture all you would have to do is 'speak' out your wish to your gadget in any which way you want, the way you would to another human. Meaning: bye-bye to QWERTY and touchpads! Far more significantly, it means that you need not even be literate to send out a message or a mail, a development with unimaginable social implications fora country like India where illiteracy is still shamefully high. However, what about the cost Says Arvind: "As neural becomes normal, it would be available across gadgets, not just on high-end smart phones and become as ubiquitous as the mobile phone itself."
Does this mean that the next generation of gadgets would come without keypads According to experts this is unlikely, simply because writing is as natural as speaking to humans! Writing at the end of the day is a more deliberate and cerebrally intense exercise that feeds on the key human faculty of thought. It would, therefore, be always an important mode of expression. And "just as speaking and writing coexist on the physical plane, so they would in the digital universe," concludes Arvind.
(c) 2013 @ 2013 DILIGENT MEDIA CORPORATION LTD. ALL RIGHTS RESERVED
[ Back To TMCnet.com's Homepage ]