Abstract

In the not too distant future, we envision information gathering agents that will have access to a large set of diverse information sources. These information sources will not belong to the agent, rather they will be information resources that are made available across the network (possibly for a fee). The agents will not maintain any real data, they will only have descriptions of the contents of these information sources. An agent will have a domain model of its area of expertise (eg, a class hierarchy describing the objects in it domain), and a description of an information source relates the contents of the source to the domain model of the agent. An agent will serve as a mediator for processing information requests, sending requests to the appropriate information sources and possibly processing the intermediate data, thus freeing a user from being aware of and sending queries directly to the information sources.

Date

September 22, 1994

Authors

Craig A Knoblock, Alon Levy

Journal

Proceedings of the Workshop on Intelligent Information Agents, Gaithersburg, MD