Casini, GStraccia, U2013-07-022013-07-022013Casini, S and Straccia, U. 2013. Defeasible inheritance-based description logics. Journal of Artificial Intelligence Research, pp 1-561076-9757http://www.cair.za.net/sites/default/files/outputs/casini%20straccia.pdfhttp://hdl.handle.net/10204/6852Copyright: 2013 Association for the Advancement of Artificial Intelligence. This is an ABSTRACT ONLY. The definitive version is published in Journal of Artificial Intelligence Research, pp 1-56Defeasible inheritance networks are a non-monotonic framework that deals with hierarchical knowledge. On the other hand, rational closure is acknowledged as a landmark of the preferential approach to non-monotonic reasoning. We will combine these two approaches and define a new non-monotonic closure operation for propositional knowledge bases that combines the advantages of both. Then we redefine such a procedure for Description Logics (DLs), a family of logics well suited to model structured information. In both cases we will provide a simple reasoning method that is built on top of the classical entailment relation and, thus, is amenable of an implementation based on existing reasoners. Eventually, we evaluate our approach on well-known landmark test examples.enDescription logicsArtificial intelligenceDefeasible inheritance networksArticleCasini, G., & Straccia, U. (2013). Defeasible inheritance-based description logics. http://hdl.handle.net/10204/6852Casini, G, and U Straccia "Defeasible inheritance-based description logics." (2013) http://hdl.handle.net/10204/6852Casini G, Straccia U. Defeasible inheritance-based description logics. 2013; http://hdl.handle.net/10204/6852.TY - Article AU - Casini, G AU - Straccia, U AB - Defeasible inheritance networks are a non-monotonic framework that deals with hierarchical knowledge. On the other hand, rational closure is acknowledged as a landmark of the preferential approach to non-monotonic reasoning. We will combine these two approaches and define a new non-monotonic closure operation for propositional knowledge bases that combines the advantages of both. Then we redefine such a procedure for Description Logics (DLs), a family of logics well suited to model structured information. In both cases we will provide a simple reasoning method that is built on top of the classical entailment relation and, thus, is amenable of an implementation based on existing reasoners. Eventually, we evaluate our approach on well-known landmark test examples. DA - 2013 DB - ResearchSpace DP - CSIR KW - Description logics KW - Artificial intelligence KW - Defeasible inheritance networks LK - https://researchspace.csir.co.za PY - 2013 SM - 1076-9757 T1 - Defeasible inheritance-based description logics TI - Defeasible inheritance-based description logics UR - http://hdl.handle.net/10204/6852 ER -