Representation and Use of Explicit Justifications for Knowledge Base Refinements

We discuss the representation and use of justification structures as an aid to knowledge base refinement We show how justifications can be used by a system to generate explanations - for its own use-of potential causes of observed failures. We discuss specific information that is usefully included in these justifications to allow the system to isolate potential faulty supporting beliefs for its rules and to effect repairs. This research is part of a larger effort to develop a Learning Apprentice System (LAS) that partially automates initial construction of a knowledge base from first-principle domain knowledge as well as knowledge base refinement during routine use. A simple implementation has been constructed that demonstrates the feasibility of building such a system.