Theoretical Foundations for Belief Revision

Belief revision systems are AI programs that deal with contradictions. They work with a knowledge base, performing reasoning from the propositions in the knowledge base, "filtering" those propositions so that only part of the knowledge base is perceived the set of propositions that are under consideration. This set of propositions is called the set of believed propositions. Typically, belief revision systems explore alternatives, make choices, explore the consequences of their choices, and compare results obtained when using different choices. If during this process a contradiction is detected, then the belief revision system will revise the knowledge base, "erasing" some propositions so that it gets rid of the contradiction. In this paper, we present a logic suitable to support belief revision systems and discuss the properties that a belief revision system based on this logic will exhibit. The system we present, SWM, differs from most of the systems developed so far in two respects: First, it is based on a logic which was developed to support belief revision systems. Second, its implementation relies on the manipulation of sets of assumptions, not justifications. The first feature allows the study of the formal properties of the system independently of its implementation, and the second one enables the system to work effectively and efficiently with inconsistent information, to switch reasoning contexts without processing overhead, and to avoid most backtracking.