
[full paper] 
Gerhard Friedrich
The generation of explanations is considered as a main asset of knowledge based systems. Explanations are exploited in communication processes in order to justify the outcome of a reasoning procedure. E.g., during a sales process a client wants to know why an offered solution (i.e., a product or service) implies particular properties. In this paper we show that current approaches of generating explanations fall short. These approaches can lead to spurious explanations with respect to a proposed (or selected) solution. We introduce an extension of current explanation principles such that all explanations of properties regarding a proposed solution are wellfounded explanations. Such wellfounded explanations can be computed by additional polynomial computation costs compared to traditional explanation algorithms.
Keywords: Constraint Satisfaction, Explanation
Citation: Gerhard Friedrich: Elimination of spurious explanations. In R.López de Mántaras and L.Saitta (eds.): ECAI2004, Proceedings of the 16th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2004, pp.813817.