belief revision


Also found in: Wikipedia.

belief revision

(artificial intelligence)
The area of theory change in which preservation of the information in the theory to be changed plays a key role.

A fundamental issue in belief revision is how to decide what information to retract in order to maintain consistency, when the addition of a new belief to a theory would make it inconsistent. Usually, an ordering on the sentences of the theory is used to determine priorities among sentences, so that those with lower priority can be retracted. This ordering can be difficult to generate and maintain.

The postulates of the AGM Theory for Belief Revision describe minimal properties a revision process should have.

This article is provided by FOLDOC - Free Online Dictionary of Computing (foldoc.org)
References in periodicals archive ?
These then feed into an account of belief revision, that is, the revision of material incompatibilities in belief sets.
The contributions that make up the main body of the text are devoted to negativity bias in investorsAE reactions to board of directorsAE risk oversight disclosure; the effects of Guanxi and compensation structure on the objectivity of Chinese internal auditors; variance reporting, belief revision, and profitability, and a wide variety of other related subjects.
The effects of audit task on evidence integration and belief revision. Behavioural Research in Accounting, 6, 21-42.
One area that comes to mind is some type of probabilistic belief revision.
In the process that R updates its own mental attitudes with input variables, the problem of belief revision is involved.
Topics discussed include representing case variations for learning general and specific adaptation rules, a theorem prover with dependent types of reasoning about actions, probabilistic association rules for item-based recommender systems, semantics for containment belief revision in the case of consistent complete theories, integrating individual and social intelligence into module-based agents without a central coordinator, improving batch reinforcement learning performance through transfer of samples, managing risk in recurrent auctions for robust resource allocation, and domain-dependent view of multiple robots path planning.
The original belief revision model (Catena et al., 1998) explaining this effect, proposed an anchoring and adjustment algorithm based upon the action of two serial mechanisms during the process of learning the causal relationships between a single causal cue and an effect (see also Appendix).
Before believe changes the belief status of a proposition p, it performs a limited form of belief revision (Alchourron, Gardenfors, and Makinson 1985).
(6) Terms like "knowledge revision," "theory change," "theory revision," and "belief revision" are used as synonyms in literature.
These findings add new-evidence to the growing body of data showing that human causal learning depends on the action of several mechanisms, as proposed by the Belief Revision Model.