UCT CS Research Document Archive

Rational Defeasible Reasoning for Description Logics

Britz, Katarina, Giovanni Casini, Thomas Meyer, Kody Moodley, Uli Sattler and Ivan Varzinczak (2018) Rational Defeasible Reasoning for Description Logics. Technical Report CS18-01-00, Department of Computer Science, University of Cape Town.

Full text available as:
PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract

In this paper, we extend description logics (DLs) with non-monotonic reasoning fea- tures. We start by investigating a notion of defeasible subsumption in the spirit of defeasible conditionals as studied by Kraus and colleagues in the propositional case. In particular, we consider a natural and intuitive semantics for defeasible subsumption, and we investi- gate syntactic properties (à la Gentzen) for both preferential and rational subsumptions and prove representation results for the description logic ALC. Such representation results pave the way for more effective decision procedures for defeasible reasoning in DLs. We analyse the problem of non-monotonic reasoning in DL at the level of entailment for both TBox and ABox reasoning, and present an adaptation of rational closure for the DL en- vironment. Importantly, we also show that computing it can be reduced to classical ALC entailment. One of the stumbling blocks to evaluating performance scalability of rational closure is the absence of naturally occurring DL-based ontologies with defeasible features. We overcome this barrier by devising an approach to introduce defeasible subsumption into classical real-world ontologies. Such semi-natural defeasible ontologies, together with a purely artificial set, are used to test our rational closure algorithms. We found that performance is scalable on the whole with no major bottlenecks.

EPrint Type:Departmental Technical Report
Subjects:I Computing Methodologies: I.2 ARTIFICIAL INTELLIGENCE
ID Code:1246
Deposited By:Meyer, Thomas
Deposited On:09 Febuary 2018