An Expert Information Filtering Service for the Internet

As a National information resource, the Internet makes available vast amounts of information to a broad range of consumers. For most users, however, only a tiny fraction of the aggregate information is significant, and the overall resource only has value if that fraction can be located reliably, quickly, and automatically. Today's tools are incapable of meeting users' needs - they are too slow and require too much manual intervention. Suitable tools can be based on the use of expert database technology in the context of software for rapid filtering of evolving data streams. This project addresses one such tool: an eXpert INformation Filter (XINF) that will cull the articles in on-line information sources to provide users with those best matching their interests. The XINF will use AI techniques to adpatively refine its knowlege of users' interests based on minimal user input and feedback, so it will be simple to use and capable to high selectivity. In addition, the XINF will be implemented using scalable computing technology, to make it indefinitely expandable - as the number of users or sources increases, the XINF will simply run in parallel on a larger number of networked computers to achieve the desired level of responsiveness. Anticipated Benefits: The technology developed in this project will be broadly applicable to one of the nation's most pressing problems: how to process and extract knowledge from the vast amount of data that is produced by numerous sources every day. Thus it can serve as the basis not only for a commercial version of te XINF service, but also for numerous other fusion applications in domains as diverse as manufacturing control, real-time data analysis, and transaction processing. For that reason, it has significant potential to enhance US industrial competitiveness and national security