MYCIN and NEOMYCIN: two approaches to generating explanations in rule-based expert systems.

The prototypical rule-based expert system is MYCIN, a computer program developed in the 1970's to diagnose and recommend therapy for serious infections. MYCIN is able to explain its reasoning at any point in a consultation by listing the rules it has under consideration at that moment. However, when MYCIN's rules were used as the subject matter for a computerized infectious disease tutoring system, it became apparent that these rules contained implicit knowledge about how to perform diagnostic tasks and that this knowledge was inaccessible to the explanation system and, therefore, to students. This paper briefly describes NEOMYCIN, an expert system that makes this implicit knowledge explicit, and shows the effect that this reconfiguration of knowledge has on generating explanations.