My research interests still focus on information retrieval, with emphases on both better understanding the theoretical foundations of ret rieval and producing improved techniques for retrieving information. I have hosted information retrieval seminars involving both leading researchers and practitioners.
Among the techniques I have explored in my journal writings are: genetic algorithms to improve the effectiveness with which one can retrieve relevant -- and filter out nonrelevant -- documents; influence diagrams (Bayesian Networks) as a representation suggesting relationships among concepts and an inquirer's underlying goals for ret rieval; and best-first searching as a means of guiding probabilistic retrieval. My theoretical work has investigated: the underlying assumptions behind probabilistic retrieval models of information retrieval; the requirement that information retrieval systems embrace principles of adaptation; and the use of "structured" methods of retrieval that provide additional "meaning" in representing document's and users' needs in comparison to traditional, keyword-based methods.
My most recent work has been on two fronts: First, I have been studying how information retrieval methods can be used to support literature-based discoveries. Research has already confirmed that the published literature occasionally contains significant, but unknown, connections am ong ideas because any given, specialized literature contains only part of the connection. So my work is aimed at devising computational methods to make these connections more apparent. My other recent effort has been on describing with a formal language the ways people use documents. Here, the idea is that one can automatically capture in a factually accurate way information concerning a document's history of use, as well as its purpose and any notable associations to other documents, and that these fa cts can augment ordinary information retrieval capabilities.
My teaching interests center on information retrieval theory and practice, expert systems, and information systems development. I am the founder and leader of the University of Michigan Info rmation Systems Executive Forum for Improving Information Management, which brings Information Systems executives to Ann Arbor for discussions and presentations on various topical issues in the field of information systems. I have consulted on informatio n retrieval, business process reengineering, and information systems analysis and design for various organizations.
Michael D. Gordon