Laboratory for Intelligent Internet Research

The Internet and the Web are perceived to have a potential impact comparable to the telephone and the printing press. A salient feature of this populist hypermedia is its universality: the Web is so valuable because it thrives from contributions in terms of technology and content from around the world. Despite its value, the Web s utility is limited in various ways: content-and user-neutral transport protocols, lack of multilingual support, primitive keyword-based search interfaces, and little support for exploiting personal and community profiles in query processing. The Laboratory for Intelligent Internet Research, a partnership between IIT Bombay and Tata Consultancy Services (through its corporate R&D center Tata Research, Development and Design Center, TRDDC), is uniquely poised to attack these limitations with the Web and thereby achieve leadership in Internet content management technology. Our goal is to develop cutting-edge research ideas into papers, patents and products. To this end, as detailed in this report, we have already made considerable progress towards making Internet more intelligent along several interrelated dimensions: • Integration between computational linguistics and statistical pattern recognition • Advanced Web architectures to support the spontaneous formation of communities • Advanced Web architectures for content-cognizant information dissemination More specifically, we are working on intelligent creation and management of hypermedia, discovering and nurturing topic-based Web communities, intelligent data dissemination, knowledge extraction from hypertext documents, and search based on linguistic analyses, all of which are of great research and commercial interest. This progress report summarizes the work done in the laboratory starting from its inception up to the end of May of 2001. During this period, our research has led to publications in many high-profile conferences devoted to research on the The practical implications of our work are manifest in the software developed to demonstrate the efficacy of our research ideas. The interest garnered by our work thus far from the international research as well as practicing community indicates that tremendous opportunities lie ahead for harnessing the IP generated by the Laboratory. At the same time, exciting challenges present themselves when one examines our vision vis a vis the state of the art and practice. Our work has just begun! 2 Our Vision • One way the Internet will appear intelligent is if it can interact with the human user through a medium the user is most familiar with. This is the medium of language and image. Could we type complete sentences or …