The InfoLab Group develops intelligent interactive software systems that help people access information and solve problems on human terms. We believe that natural language is the best information access mechanism for humans---our research focuses on building software systems that respond to users' queries posed in everyday language with "just the right information."

At the heart of this effort is the START Natural Language System and the Omnibase virtual database developed by Boris Katz and his associates. START and Omnibase provide access to text, Web pages, images, and other multimedia items in response to English questions. Our combined system works by using a technique called "natural language annotation," in which retrievable information segments are associated with natural language sentences and phrases -- annotations -- that describe their contents. When a user asks a natural language question, it is first parsed into a representation we called "ternary expressions." These parsed structures are then matched against ternary expressions derived from the annotations: if a match occurs, the corresponding information is returned to the user as the answer.

An important feature of the annotation concept is that any information segment can be annotated: not only text, but also images, multimedia, and even procedures! We have even annotated database queries, which provide our system seamless access to dozens of databases on the World Wide Web.

Since it came on-line in December, 1993, START has engaged in exchanges with hundreds of thousands of users all over the world, supplying them with useful knowledge. Currently, our system answers millions of natural language questions about places (e.g., cities, countries, lakes, coordinates, weather, maps, demographics, political and economic systems), movies (e.g., titles, actors, directors), people (e.g., birth dates, biographies), dictionary definitions, and much, much more.

For more information, please refer to our publications.