Expert Humans and Expert Systems: Toward a Unity of Uncertain Reasoning

Expert systems use a variety of techniques to reason about uncertain, incomplete, or unclear knowledge. This paper considers some methods expert systems have used to attempt to match human reasoning with and about uncertainty. It discusses MYCIN certainty factors, probability and Bayes' Theorem, Dempster-Shafer belief functions, and fuzzy logic. For each method, the paper suggests how conceptual graphs can be used to implement those representations, using concepts, relations, schemata, and actors.