PARTY: A mobile system for efficiently assessing the probability of extensions in a debate

Abstract

In this paper we propose PARTY, a Probabilistic Abstract aRgumenTation sY stem that assesses the probability that a set of arguments is an extension according to a semantics. PARTY deals with five popular semantics, i.e., admissible, stable, complete, grounded, and preferred: it implements polynomial algorithms for computing the probability of the extensions for admissible and stable semantics and it implements an efficient Monte-Carlo simulation algorithm for estimating the probability of the extensions for the other semantics, which have been shown to be intractable in [19,20]. The experimental evaluation shows that PARTY is more efficient than the state-of-the art approaches and that it can be profitable executed on devices having reduced computational resources.