What Is Believed Is What Is Explained (Sometimes)

Abstract

Abstract This paper presents a formal and computa-tional methodology for incorporation of new knowledge into knowledge bases about actions and changes. We employ Gelfond and Lifschitz ’ action description lan-guage A to describe domains of actions. The knowl-edge bases on domains of actions are defined and ob-tained by a new translation from domain descriptions in J into abductive normal logic programs, where a time dimension is incorporated. The knowledge bases are shown to be both sound and complete with respect to their domain descriptions. In particular, we pro-pose a possible causes approach (PCA) to belief update based on the slogan: What is believed is what is ex-plained. A possible cause of new knowledge consists of abduced occurrences of actions and value propositions about the initial state of the domain of actions, that would allow to derive the new knowledge. 1Ve show how to compute possible causes with abductive logic programming, and present some techniques to improve search efficiency. We use examples to compare our pos-sible causes approach with Ginsberg’s possible worlds approach (P WA) and Winslett’s possible models ap-proach (PMA).