A unified approach to default reasoning and belief revision

Abstract

We present a unified approach to three important areas of AI: nonmonotonic logics, belief revision theories, and belief revision systems. The nonmonotonic logic we present, SWMC, is appropriate to support belief revision systems, because it keeps a track of dependencies between formulas. Another distinguishing feature of SWMC is the distinction between what follows soundly from a set of premises, and what can plausibly be concluded. In what concerns belief revision theories, we present a theory based on a nonmonotonic logic, SWMC, which, to the best of our knowledge, is a novel approach. Finally, we use an implementation of a belief revision system based on the logic and the belief revision theory, to illustrate these formalisms.KeywordsPartial OrderBelief RevisionDefault RuleUnderlying LogicDefault ReasoningThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.