Research on Deductive Inference for Large Data Bases

Abstract : The research has as its major goal the construction of software tools to aid on-line decision makers and data base users in accessing information relevant to their needs, in understanding the full data base search implications of their requests, and in reviewing and evaluating the utility of derived answers. The conceptual framework within which this research has been carried out is based upon mathematical logic. It is becoming increasingly clear that logic is, highly relevant not only to reasoning about data but to query language design, to data structuring, to the support of high level user views, to maintaining the integrity of data bases, and to making the transition from present day data-based systems to future knowledge-based systems. The main software tool that has been implemented as part of this research is called DADM (for Deductively Augmented Data Management). This report describes the design, implementation, and current capability of this prototype system. DADM adds a general knowledge base and a deductive processor to a data management system. These components are used to control the creation of intelligent data base access strategies and the construction of evidence to support derived answers.