Rule Refinement by Domain Experts in Complex Knowledge Bases

We research how subject matter experts can develop knowledge-based systems that incorporate their expertise. Our approach is to develop a learning and problem solving agent, called Disciple, that an expert can teach by explaining it how to solve specific problems, and by critiquing its attempts to solve new problems (Tecuci 1998). The knowledge base of the agent is structured into an object ontology that contains a hierarchical description of the objects and features from an application domain, and a set of task reduction rules expressed with these objects. The Disciple approach has already been applied to develop knowledge-based agents for complex military tasks such as course of action critiquing and center of gravity analysis (Tecuci et al. 2002). In this paper we present a new integrated approach to support a domain expert in refining the rules from an agent’s large knowledge base.