Introducing probabilistic reasoning in defeasible argumentation using labeled deductive systems

Abstract

LabeledDeductive Systems (LDS)were developed as a rigorous but fexiblemethodology to formalize complex logical systems, such as temporal logics, database query languages and defeasible reasoning systems. LDSAR is a LDS-based framework for defeasible argumentation which subsumes diferent existing argumentation frameworks, providing a testbed for studying diferent relevant features (such as emerging logical properties, ontological aspects, semantic characterization, etc.) This paper discusses some relevant issues concerning the introduction of probabilistic reasoning into defeasible argumentation. In particular, we consider a first approach for recasting the existing LDSAR framework in order to incorporate numeric attributes (certainty factors) as part of the argumentation process.