![]() On the logic of theory change: Partial meet contraction and revision functions. Various examples are given to illustrate the approach, including Horn clause revision, revision in extended logic programs, and belief revision in a very basic logic called literal revision. As well, iterated revision is addressed and the Darwiche/Pearl postulates are shown to be compatible with our approach. Crucially, both of these additions are redundant in the original AGM framework, and so we extend, rather than modify, the AGM approach. To obtain the representation result, we add a new postulate to the AGM postulates, and we add a constraint to preorders on worlds. The classical AGM postulates are expressed in this framework and a representation result is established between the postulate set and certain preorders on possible worlds. We show that AGM-style revision can be obtained even when extremely little is assumed of the underlying language and its semantics in fact, one requires little more than a language with sentences that are satisfied at models, or possible worlds. In this article, we consider the question of what the minimal requirements are on a logic, such that the AGM approach to revision may be formulated. ![]() This is a significant limitation, since many representation schemes in AI don’t subsume propositional logic. ![]() The standard (AGM) approach to belief revision assumes that the underlying logic contains classical propositional logic. In artificial intelligence, a key question concerns how an agent may rationally revise its beliefs in light of new information.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |