Towards an Ontology-Independent Problem Solver

Questions posed to knowledge based systems typically omit information that is essential for answering them. Problem solvers, therefore, take on the responsibility of elaborating them before they can be answered. Automating this process is especially challenging when the problem solver is meant to work with a variety of knowledge bases. We have studied this problem in the context of a system used to answer questions posed by naive users who are unfamiliar with the content and organization (i.e. the ontology) of the knowledge bases. This separation between the builders and the users of the knowledge bases requires the problem solving system to elaborate the questions automatically with information drawn from the knowledge base being queried. Our approach is shown to work well on different knowledge bases and it answers a broad range of unanticipated questions in the physics domain.