Information aggregation and agent interaction patterns in InfoSleuth

The MCC InfoSleuth/sup TM/ Project is an agent-based system for information gathering and analysis tasks performed over networks of autonomous information sources. A key motivation of the InfoSleuth system is that real information gathering applications require long-running monitoring and integration of information at various levels of abstraction. To this end, InfoSleuth agents enable a loose integration of technologies allowing: (1) extraction of semantic concepts from autonomous information sources; (2) registration and integration of semantically annotated information from diverse sources; and (3) temporal monitoring, information routing, and identification of trends appearing across sources in the information network. In this paper we discuss the agents in InfoSleuth applications and the goal-driven interaction patterns that enable them to dynamically organize and cooperate to perform integrated and temporal information-gathering tasks.