Improving WWW Information Gathering: More Features and More Personal

Today, the typical user enters a two word query to Google to satisfy his/her information need on the WWW and then browses a miniscule subset of the millions of results. Many enhancements to current search technology are on the horizon that will change how the WWW is used as an information resource. In this talk, I will discuss research on two enhancements: methods for personalizing search and extending the features used to characterize documents and queries. SurfAgent learns user information interests to support a range of information gathering services. Collaborative filtering can leverage shared information interests among a community. Document features such as the type or genre of the document and the imagery contained in it complement the currently used document information such as terms and links. Adele Howe is a Professor of Computer Science at Colorado State University. She received her Ph.D. in 1993 from the University of Massachusetts in the area of AI planning. Her current research interests include AI planning and scheduling,evaluating AI systems,and web-based information agents.