The Spoken Web is a voice-based equivalent of the World Wide Web, primarily designed for semi-literate populations of developing countries. Users can dial into the Spoken Web through their mobile or regular phones, create their own voicesites (analogous to websites) in local languages, and browse voicesites of others. Voicesites can be created to help people seek expert advice on agriculture or health or other important activities, and even conduct financial transactions. The vision is to create an information ecosystem that helps provide access to Internet-like information services, but primarily through voice, to make the medium accessible to the (yet) underserved bottom-of-the-pyramid population.
More information about the Spoken Web project can be found here:
http://en.wikipedia.org/wiki/World_Wide_Telecom_Web
http://domino.research.ibm.com/comm/research_projects.nsf/pages/pyrmeait.index.html
The IBM India Research Lab has been working on the Spoken Web project since the last two years. We have the following internship openings for the summer of 2009. Please email your CV and area of interest by March 25th, 2009.
- Content search on the Spoken Web:
One of the foremost challenges with voice is to make it searchable. The search could be performed on meta-content, could be parameterized based upon the VoiceSite templates or it could be done over raw speech data. However, current state-of-the-art speech recognition technologies offer only limited help, especially given the multitude of languages and dialects spoken in countries like India. A different direction is to create a multimodal interface that allows semiautomatic categorization or textual annotations of voice clippings, while ensuring that the simplicity of voice-based interaction is retained for the end-user. For this project, we are therefore looking for candidates with a background in information retrieval, but who can apply IR techniques in innovative ways suited for the social context in which the Spoken Web is being used.
- Scalability of the Spoken Web technology:
To ensure the widest possible dissemination of the Spoken Web technology, it must be low-cost, administratively easy to install, and extremely robust, so that non-profits and community organizations can easily set up and maintain systems on their own. These goals can be achieved by migration of the hardware to low-cost off-the-shelf components, analysis and improvement of server performance, and innovations on tools and processes for remote repairs and troubleshooting of the software and hardware. For this project, we are therefore looking for candidates who have a background on system building and performance evaluation.
We are extremely excited about the impact and potential of the Spoken Web. We look forward to hear from you and share our enthusiasm. The working environment at IRL motivates innovation; we are sure that you will have a lot of fun here.
Back to Internships Information
Back to Scholarships Information