A Flexible Framework for Uncertain Expertise

In this paper we argue that the development of knowledgebased systems built to work in partially uncertain domains bene t from the use of di erent conceptualisations for certain and uncertain parts of the knowledge. We present conceptualisations that have proven to be useful, namely the KADS model of expertise and a causal model of uncertainty that re ects well known approaches to uncertain reasoning like Bayesian belief nets. After a brief introduction to these conceptualisations we propose a translation approach that aims at an integration of these conceptualisations in a common knowledge model that can be used in a knowledge engineering process.