A preferential framework for trivialization-resistant reasoning with inconsistent information

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Paraconsistent entailments based on more than two truth-values are useful formalisms for handling inconsistent information in large knowledge bases. However, such entailments suffer from two major drawbacks: they are often too cautious to allow intuitive classical inference, and are trivialization-prone. Two preferential mechanisms have been proposed to deal with these two problems, but they are formulated in different terms, and are hard to combine. This paper is a step towards a systematization and generalization of these approaches. We define an abstract framework, which allows for incorporating various preferential criteria into paraconsistent entailments in a modular way. We show that many natural cases of previously studied entailments can be simulated within this framework. Its usefulness is also demonstrated using a concrete domain related to ancient geography.

Original languageEnglish
Title of host publicationLogics in Artificial Intelligence - 13th European Conference, JELIA 2012, Proceedings
Pages463-475
Number of pages13
DOIs
StatePublished - 2012
Externally publishedYes
Event13th European Conference on Logics in Artificial Intelligence, JELIA 2012 - Toulouse, France
Duration: 26 Sep 201228 Sep 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7519 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference13th European Conference on Logics in Artificial Intelligence, JELIA 2012
Country/TerritoryFrance
CityToulouse
Period26/09/1228/09/12

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'A preferential framework for trivialization-resistant reasoning with inconsistent information'. Together they form a unique fingerprint.

Cite this