Integrating machine-learning techniques in knowledge-based systems verification

Abstract

A significant problem in the development of Knowledge-Based Systems (KBS) is its verification step. This paper describes an expert system verification approach that considers system specifications, and consequently, Knowledge Bases (KB) to be partially described when development starts. This partial description is not necessarily perfect and our work aims at using Machine Learning techniques to progressively improve the quality of expert system Knowledge Bases. by coping with two major KB anomalies: incompleteness and incorrectness. In agreement with the current tendency, KBs considered in our approach are expressed in different formalisms. Results obtained with two different learning algorithms, confirm the hypothesis that integrating machine learning techniques in the verification step of a Knowledge-Based System life cycle, is a promising approach.