Incorporating Specificity in Extended Logic Programs for Belief Revision

In this papera newoperatorfor agent's implicit belief revisionis presented.Thisoperatoris conceptually based on the following requirement:anagent should find first anexplanationfor the observationsit makes beforeit tries to revise its beliefs. Asa consequence, the proposedbelief revision modeldoesnot agreewith the principle of minimalchange, commonlyaccepted. A translation of anagent's knowledgebase, expressed by meansof a conditional defaults set for an abducrive extendedlogic program,is also presented. This translation allows us to makeexplicit the specificity that exists amongdefaults andalso the logic program allows us to determinethe possible explanationsfor a recently observedfact.

• Type: Conference
• Year: 1998
• Where: FLAIRS
• Authors: Berilhes Borges Garcia, José Gabriel Pereira Lopes
• Keywords: Artificial Intelligence, Asa Consequence, Belief Revisionis Presented.Thisoperatoris, FLAIRS 1998, Proposedbelief Revision Modeldoesnot