Using Conceptual Graphs to Represent Database Inference Security Analysis

Abstract

This paper describes an approach to database inference analysis based on conceptual graphs. The database inference problem is briefly described. Previous approaches are summarized, followed by a presentation of our inference model, called AERIE. The notions of an inference target class and an inference method class are introduced with examples given. Conceptual graphs are introduced as our means of representing database inference knowledge, as a first step toward analyzing and detecting database inference problems. The classification of inference target classes and the use of conceptual graphs for database inference detection are two important contributions of this paper. Four examples are used to illus trate the approach. We discuss some interesting issues raised by this work, and offer conclusions and our plans for future research.