Jo ao P. Martins and Stuart C. Shapiro, A model for belief revision, Artificial Intelligence 35 (1988), 25--79.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....communities, as we show in [31, 38] Obviously, our objective is not to propose a general framework for multi agent belief revision. We will be satis ed if we can nd a very simple solution that could enables us to illustrate the e ects of the social reasoning mechanism on belief revision. Like [25, 16], we consider that a procedure of belief revision is composed of the following phases: 1. detection: an inconsistency in the system is detected; 2. identi cation: the culprit(s) of the inconsistency are found; 3. decision: a context (a subset of consistent beliefs) to be maintained is chosen; ....

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Articial Intelligence, 35(2), 1988.

.... modal logics with the addition of operators such as K a OE ( a knows that OE ) and B a OE ( a believes that OE ) In AI, research in the area of epistemic logic has concentrated on three main problems: the relationships between knowledge and belief and between truth and justification ( Had91b, MS88, Bou92] consistency and logical omniscience ( Had88] and nesting of epistemic operators as in X believes that Y knows Z, Had91b] The problem of quantified epistemic logics has not been well researched in AI, though it does appear in [Had91b] Here two belief revision models are briefly ....

.... between truth, knowledge, belief, inference, questioning a belief and retraction, without being so concerned with computational tractability [Had91b] while the second was designed as a computational model, dealing with a subset of these issues, and has been implemented in the SNeBR system [MS88] 4.1 The Nested Intensional Model One major advantage of = FOL over First Order Logic (FOL) is its ability to handle nested operators. Nested epistemic operators are necessary if there is more than one agent in a system, since it then becomes necessary to be able to express such statements as ....

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Artificial Intelligence, 35(1):25--79, 1988.

....context, as modeled in typical AI applications, we usually want reason maintenance as e.g. in the toy example given above. Summarizing, we see that belief revision and reason maintenance are not genuinely connected with each other, as it sometimes seems to be perceived in AI (cf. Martins and Shapiro, 1988 ] However, as will be shown, it is not necessary to add reason maintenance as a primitive notion to a theory of belief revision. Reason maintenance will result as a side effect when we choose the right contraction operation. 5 Contracting Finite Bases As spelled out in the previous ....

Jo~ao P. Martins and Stuart C. Shapiro. A Model for Belief Revision. Artificial Intelligence 35(1): 25--79, May 1988.

....a support. In an assumption based truth maintenance system (ATMS) which is a kind of belief revision system, the support of each proposition contains hypotheses (nonderived propositions) that produced it. In the SWM system [ Martins, 1983, CHAPTER 3. LEARNING AT THE RULE SELECTION LEVEL 44 Martins and Shapiro, 1988 ] this support is named an origin set. An origin set of a proposition contains every hypothesis used in its derivation. Using the origin set, when a contradiction is detected, we should be able to identify exactly which assumptions were used in the derivation of the contradictory propositions. ....

.... disjunction, negation, implication, universal quantifier, and existential quantifier [ Shapiro, 1979c ] They are designed to provide closeness to human reasoning, structural simplicity, and expressibility in the areas of natural language understanding, knowledge representation, and reasoning [ Martins and Shapiro, 1988, Shapiro, 1979a, Shapiro, 1979b ] Some of the connectives and quantifiers are briefly introduced below. And or is symbolized as n WV j i , and the formula n WV j i (A 1 ; A 2 ; A n ) is true when at least i and at most j of the n arguments are true. And or generalizes conjunction, ....

[Article contains additional citation context not shown here]

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Artificial Intelligence, 35:25--79, 1988.

....cognitive model of belief revision. Area(s) belief revision, explanatory coherence, connectionism, nonmonotonicity Presentation Format: Talk Or Poster Explanatory Coherence as a Model for Belief Revision Christina Carrick Simon Fraser University Burnaby, BC V5A 1S6 1 Introduction In [MS88], Martins and Shapiro describe several specific tasks which must be solved in order for a system to successfully revise beliefs in the face of contradictory evidence: inference, dependency recording, nonmonotonicity, and disbelief propagation. Inference is concerned with how new beliefs can be ....

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Artificial Intelligence, 35(1):25--79, 1988.

....semantics. Each propositional node has a support that is used to decide if a node is a part of the current context or not. The support of a proposition can be used to build a proof for the proposition. The support of a proposition contains the assumptions used to prove the proposition. SWM[MS88] introduced the concept of supported wffs. A supported wff is a 3 tuple P;t;a where P is a wff, t is the origin tag taking values from the set fhyp;derg, and a is the origin set. The origin set is a set and consists of all the hypotheses that are used to prove P (ATMS style) A proposition ....

....default reasoning module for this project. 4.3 Default Reasoning in SNePSwD Default Reasoning has been implemented earlier in one of the previous versions of SNePS and is called SNePSwD. SNePSwD is based on the SWMC (Shapiro, Wand, Martins and Cravo) CM92] logic which is an extension of the SWM[MS88] logic. SWMC is a nonmonotonic logic that supports ATMS like systems with reasoning capabilities. SWMC allows for default reasoning, i.e. the use of default rules, which are not universally true (rules with exceptions) SWMC allows expression of default rules and exceptions to these rules. It ....

Jo ao P. Martins and Stuart C. Shapiro, A model for belief revision, Artificial Intelligence 35 (1988), 25--79.

....complexity for retrieving and storing instances. The second implementation of the shadowing method is using the origin set in an assumption based truth maintenance system (ATMS) In ATMS, an origin set is associated with each proposition to keep track of and propagate propositional dependencies [7] Using the origin set, when a contradiction is detected, we should be able to identify exactly which assumptions were used in the derivation of the contradictory propositions. From the viewpoint of deductive learning, propositional dependencies represented by an origin set can be regarded as a ....

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Artificial Intelligence, 35:25--79, 1988.

....and there is a migrated instance hR2; fon=Rgi in I R1cq satisfying OE 1 oe fon=Rg. The second method of shadowing uses an origin set (OS) that is associated with each proposition to keep track of and propagate propositional dependencies in an assumption based truth maintenance system SNeBR [3]. In the above example, OS of R1 is fR1g and OS of R2 is fR1, trans(on)g since R2 is derived from the two propositions. From the viewpoint of deductive learning, propositional dependencies represented by OS can be regarded as a type of experience. A subset superset comparison between OSs of two ....

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Artificial Intelligence, 35:25--79, 1988.

....information. To deal with this, the system needs two facilities: 1. The ability to recognize and trap explicit contradictions so that something can be done about them. 2. The ability to retract stored information inferred from information that is later retracted. SNePS 2. 1 includes SNeBR [ Martins and Shapiro, 1988 ] a Belief Revision system that has these two abilities. When some wff is entered or inferred that directly contradicts one that is already stored, SNeBR opens a dialogue with the user: all(x) Bird(x) Flies(x) all(X) BIRD(X) FLIES(X) all(x) Penguin(x) Bird(x) all(X) PENGUIN(X) ....

Jo~ao P. Martins and Stuart C. Shapiro. A model for belief revision. Artificial Intelligence, 35:25--79, 1988.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC