The InfoSleuth Project: Intelligent Search Management via Semantic Agents

InfoSleuth is a research project at MCC that is developing and deploying new technologies for finding information available both in corporate networks and in external networks, such as networks based on the emerging National Information Infrastructure (NII). The InfoSleuth research is based on the Carnot technology that has been developed at MCC over the last 4 years. Carnot has been successfully used to integrate heterogeneous corporate information resources. Carnot executes queries in a distributed environment by dispatching autonomous computing agents to remote sites where they access databases and cooperate among themselves to properly merge resulting data into understandable information.The InfoSleuth project will investigate the use of Carnot technology in a more dynamically changing environment, such as the Internet, where new information sources are constantly being added and for which there is no formal control of the registration of new information sources. In this type of environment, traditional techniques for expressing and optimizing database queries are inadequate because of the rapidly changing schema information and the fuzzy nature of the queries. InfoSleuth will build on Carnot semantic modeling capabilities to enable “deep” descriptions of available information sources. InfoSleuth will deploy semantic agents to carry out distributed, coordinated, self-adapting search algorithms. MCC is seeking industrial sponsors to participate in the InfoSleuth research.