Research Agents and Inferential Retrieval


This SBIR (NSF) phase-I project addresses the problem of search, especially Internet based search, from a unique perspective of research agents that mediate between, and on behalf of, the user and more traditional information retrievel systems. The approach is innovative in a number of ways. It

  • helps the user in constructing and elaborating mental models of the research context that motivates the user's search behavior

  • leverages such context with fast, parallel, inferential search that filters and organizes search results from traditional search engines

  • supports critical thinking as it bears on the research issues being investigated by the user and

  • envisions research as a collaborative process in which the fruits of each researcher's labor may be published electronically as active knowledge structures that can inform and organize subsequent searches, and which can form the basis of an improved understanding of the domain.

    Direct commercial applications include information access and management on the Internet, and in vertical markets; preparing briefs for case or patent law; market analysis for traders and brokers; intelligence analysis; primary and continuing education; university-based research; and conducting and managing national research interests.

    SHRUTI serves as the fast, parallel, and evidential inference engine in this project.

    This project is funded as a subaward of an NSF, SBIR Phase I award to Cognitive Technologies Inc.