It is not easy to miss the glint in IBM Research (India) director Manish Gupta?s eyes. The Indian research lab has been chosen to lead worldwide research on mobile Web?one of the big bets or ?main focus areas? of IBM Research. ?We are the first non-US lab in IBM Research?s history to lead a worldwide big bet and we are already seeing success. It represents a commitment of $100 billion in next five years,? he exclaims. The signature project, spoken Web, is already running successful pilots. Here, voice sites are being created and linked to each other through hyperlinks just like on worldwide Web. Search capabilities on voice sites are also in the pipeline. As he leads the efforts to create a parallel Web, Gupta is equally gung-ho on their pioneering research in analysing social networking patterns and offering services on a cloud. Having led research on system software for the IBM Blue Gene supercomputer and other Deep Computing platform, Manish Gupta has over 70 papers and has filed 16 patents. In a chat with Pragati Verma, Manish Gupta offers a peep into disrupting research work at IBM India labs. Excerpts:
Indian research lab has been chosen to lead the worldwide research on mobile Web. How is it shaping up?
Each year, IBM?s research division places what it calls its big bets, which are key areas where we see some major potential for breakthroughs and leading disruptions. It represents a commitment of about $100 billion in next five years. It involves several IBM labs but this is the first time in IBM research?s history that a non-US lab was asked to lead a worldwide big bet.
Our signature project is spoken Web. Spoken Web helps people create voice sites using a simple telephone?mobile or landline. When a user wants to create a voice site, he or she needs to call a number and a software called VoiGen helps the person create a new site. The user gets a unique phone number which is analogous to a URL and when other users access this voice site, they get to hear the content uploaded there. It can transform the way information gets disseminated and services get offered, essentially over the phone network. Analogous to Web pages, you will have voice websites.
How is this different from interactive voice response (IVR) systems we have today?
In its simplest form, it could be plain recorded information. But we have developed two key things at our labs. One is we have made it very easy for a layman to answer a few simple questions over phone. Even people not familiar with IP can make a voice site, who might not know how to do an IVR site. Also, we can link one voice site to another.
The hyperlinks and the ability to seamlessly go from one page to another with a simple click is what gave the Web its power.
We have developed hyper speech transfer protocol. Here while you are listening to one voice site, you could traverse to another voice site and your call will get transferred and your application context will get transferred.
You can do things like call up a store?s voice site and order some things and get connected to a secure bank payment gateway managed by a bank. If you are not comfortable giving your credit card details to your local grocery store, we have developed a protocol for you to allow only a card company to handle those details through which money is transferred.
You can think of it as IVR on steroids. Think of the world before the Web. We could create documents but one document would not link to another. This advance will be a similar transformation.
Are you also working on taking the spoken Web out of the confines of labs to market?
In India, we have three pilots going on. For instance, farmers in Gujarat and Karnataka can post information about crops and mandi agents can post information about crops. Farmers can thus get timely information about crops and can post questions about diseases and fertilisers. We are also trying to add navigation and search capabilities. Unlike the regular Web, where things are visual, it is not easy in spoken Web.
You also work with leading telcos in the country? What are their biggest demands from tech research compnaies?
We are also working on setting up services for an enterprise to increase productivity of a mobile workforce. We have a project called snazzy to get business insights into a social networking. A typical telecom company?s Business Intelligence (BI) tool looks at customers who generate big bills. We look at their CDRs (calling detail records) and see who?s calling whom. This way, we identify communities and structures where telecom companies can offer promotions and incentives.
We can help telcos retain key connections points and identify key acquisitions targets from rival firms in these communities. We have done these with four companies in this.
As the Indian market grows, are you developing more solutions for emerging markets?
We are looking at unique problems and unique solutions that arise in countries like India. We recognise that you can?t just take technologies from Europe and the US and bring them here. Spoken Web, for instance, could have been invented in the US. Issues here are different. A delay of five seconds in accessing a website might irritate some surfers but a sugarcane farmer will be willing to wait two minutes to get information on the Web. It?s more important to give him local content in his language. We are also looking at areas like financial inclusion. Traditional bank branches won?t be enough in India and we need to replicate some of the work we have done in mobile Web area. We need solutions better than kiosks. I expect solutions to come from mobile technologies.
Do you expect spoken Web to replace the current worldwide Web ever?
I would say the two will get integrated. To begin with, we expect spoken Web to reach places where traditional Web hasn?t reached. But later, you can access information as text or voice on the worldwide Web. Then you can think of text to speech or speech to text technologies. But it will take some time before we reach that point as there are various accents and vocabularies.