Catalytic inference analysis: detecting inference threats due to knowledge discovery

Knowledge discovery in databases can be enhanced by introducing "catalytic relations" conveying external knowledge. The new information catalyzes database inference, manifesting latent channels. Catalytic inference is imprecise in nature, but the granularity of inference may be fine enough to create security compromises. Catalytic inference is computationally intensive. However, it can be automated by advanced search engines that gather and assemble knowledge from information repositories. The relentless information gathering potential of such search engines makes them formidable security threats. This paper presents a formalism for modeling and analyzing catalytic inference in "mixed" databases containing various precise, imprecise and fuzzy relations. The inference formalism is flexible and robust, and well-suited to implementation.