This paper aims to describe ongoing research into the issue of explanation in intelligent systems. Explanation has been considered as an essential facility in traditional expert systems. The existing methods do not take into consideration the linguistic context of the input question. Our approach starts from the remark that we cannot have a complete explanation if we do not take into consideration the different interpretations of the question. In this paper we discuss the problem of explanation generation in expert systems and present an environment that generates complete explanations (the 'Explanatory Environment') and an interface that allows experts to insert easily data to the Explanatory Environment (the 'Knowledge Acquisition System for Explanation').
展开▼
