The transition from data management to knowledge management

Research in Data Management and Artificial Intelligence has led to systems capable of efficiently searching vast quantities of data and to systems that use encoded expertise to solve problems in an "intelligent" fashion. This paper describes recent progress in developing systems that incorporate application specific expertise and apply that expertise in reasoning with and about data stored within a conventional data management system (DMS). We call such systems Knowledge Management Systems and we discuss the features and capabilities of a prototype Knowledge Manager: KM-1. KM-1 employs a logic-based "reasoning engine" (deductive processor) to derive implicit information from the explicit data stored within a "searching engine" (a relational data management system). Both rule based expertise and user queries are expressed in an Englishlike canonical form of first order predicate logic. Data access plans, derived from queries, and evidence chains that explain answers are presented to the user in an easy to interpret graphical form. The paper concludes with a brief discussion of current research in Logic Based Systems, AI, and database technology that is relevant to the development of future knowledge management systems.