Using maximum entropy in a defeasible logic with probabilistic semantics

James Cussens, Anthony Hunter

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

1 Citation (Scopus)

Abstract

In this paper we make defeasible inferences from conditional probabilities using the Principle of Total Evidence. This gives a logic that is a simple extension of the axiomatization of probabilistic logic as defined by Halpern's AX 1. For our consequence relation, the reasoning is further justified by an assumption of the typicality of individuals mentioned in the data. For databases which do not determine a unique probability distribution, we select by default the distribution with Maximum Entropy. We situate this logic in the context of preferred models semantics.
Original languageEnglish
Title of host publicationProceedings of the 4th International Conference on Information Processing and Management of Uncertainty in Knowledge-Based Systems (IPMU-92)
PublisherSpringer
Pages43-52
Volume682
DOIs
Publication statusPublished - Jul 1992

Publication series

NameLecture Notes in Computer Science

Fingerprint

Dive into the research topics of 'Using maximum entropy in a defeasible logic with probabilistic semantics'. Together they form a unique fingerprint.

Cite this