We will present an overview of qualitative probabilistic networks in the context of skin diseases with children. The basic framework will be explained, with particular attention to the notions of qualitative influence, product synergy and intercausal reasoning. The drawbacks of this basic approach soon become apparent; most notably, the problem of overabstraction or overgeneralization. Answering the need for refinement, we discuss the formalism of enhanced qualitative probabilistic networks. Finally, we will propose to include the noisy-MAX concept from Bayesian networks into the qualitative scheme. To this end, we provide a proof for the qualitative product synergy the noisy-MAX displays and show how it can function as a guideline to incorporate multi-valued variables in the enhanced qualitative paradigm.

Explanation of reasoning in expert systems is necessary for debugging the knowledge base, for facilitating their acceptance by human users, and for using them as tutoring systems. Influence diagrams have proved to be effective tools for building decision-support systems, but explanation of their reasoning is difficult, because inference in probabilistic graphical models seems to have little relation with human thinking. The current paper describes some explanation capabilities for influence diagrams and how they have been implemented in Elvira, a public software tool. 1