Planning-Based Information Agents
Abstract
Networked information systems are making so much data available that people can't find it themselves. Software agent technology promises to amplify human decision making capabilities by gathering information from disparate sources in parallel and integrating it in real time. However, in order to make today's prototype systems realize their potential, several bottlenecks must be overcome. First, information-gathering agents need robust and efficient execution so they can process large data sets, cope with network failure and site outage. Secondly, in order to scale to the level of thousands of information sources, agents need algorithms for locating information sources; automatically creating wrappers for those sources, processing XML based representations of those sources, and routing queries to the appropriate sources.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2001
- Accession Number
- ADA391739
Entities
People
- Daniel S. Weld
Organizations
- University of Washington