Angry with a friend who is too busy to pick up the phone? Just leave a voice message and he will take the hint through an on-screen 'emotica' -- a new age technological marvel, to hit the market soon.
'Emotive Alerts,' a pathbreaking research on voicemail systems undertaken by the Massachusetts Institute of Technology, promises to revolutionise the cellular phone technology, where labelling of messages will be possible using a software that identifies the tone of a callers' voice and tags the message as 'urgent', 'happy', 'formal', 'informal' or 'sad', converting them into popular emoticas.
On receiving such an emotica, for example, a 'smiley', the subscriber can prioritise his messages, simplifying the common problems of a jammed voicemail box and missing important calls.
He can skip the cumbersome process of having to go through the entire voicemail content to sort out the urgent ones.
One can also start the day on a happy note by sorting the 'happy' messages first, say MIT researchers Zeynep Inanoglu and Ron Caneel of the Media Lab who designed it and recently published in the science journal Newscientist.com.
According to the research, the first of its kind, "the first ten seconds of voicemail messages were used for labelling and training."
Based on a study by scientists Whittaker, Hirshbeg et al, (All Talk and All Action) which claims that human beings choose in the first 10 seconds whether to stop or keep listening to the voice, 361 message labels were compared with a previously stored data bank of hundreds of voice messages according to their 'pitch,' 'loudness' and 'rhythm.'
By extracting 'salient acoustic features' from various voicemails, the researchers decided to consider four 'axes' in voicemail correspondence. They have categorised these axes as 'valence' (happy, sad) 'activation' (calm, excited) urgency, formality and informality, the researchers say.
During trial runs on real messages, the software was actually able to distinguish between emotions ranging from excited and calm, happy and sad, but faltered during the formal and informal and urgent and non-urgent tones.
However, when the product is ready, it may take some time to reach the Indian subscribers. Not much is being done to update and popularise the cellular telephony market here.
"The closest that the country has reached vis-à-vis research on artificial intelligence are works on speech simulations by the Bangalore-based organisation Simputer," says Dr Amitabh Mukherjee, faculty of the Computer Science and Engineering department of the Indian Institute of Technology, Kanpur.
"As far as I know no similar research is going on here in this country and it can be attributed to the fact that the telephony penetration here is so low," he says.
"This is clearly a market driven research. The end users in India are very low compared to the western market to warrant the popularity of a such a software," Mukherjee says.
"In the artificial intelligence sector, the Simputer group has been developing voice recognition software which includes the encoding of human speech, which is the nearest we have gone to this kind of research," he says.
The lack of 'enough end customers' coupled with a still growing communications market, is enough reason that such research would take a long time to reach the Indian consumers, he says.
