Deciding Subsumption in Defeasible ELI with Typicality Models
LOGICS IN ARTIFICIAL INTELLIGENCE, JELIA 2023(2023)
摘要
Some reasoning methods for Defeasible Description Logics (DDLs) suffer from quantification neglect (QN) as they omit un-defeated information for quantified objects. Reasoning in defeasible EL perpendicular to based on so-called typicality models (TMs), which extend canonical models of classical EL perpendicular to, can alleviate QN. The DDL ELI perpendicular to extends EL perpendicular to by inverse roles, i.e., a limited form of value restriction. Extending TMs to inverse roles is challenging due to their interaction with existential restrictions. In this paper, we develop TMs for ELI perpendicular to for 4 different semantics reliant on rational and relevant closure. Our computation methods for those TMs are effective decision procedures for subsumption in defeasible ELI perpendicular to and the stronger forms of TMs can mitigate QN.
更多查看译文
关键词
Description Logics,Defeasible Logics,Nonmonotonic Reasoning
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要