Combining Extension-based Semantics and Ranking-based Semantics for Abstract Argumentation

Elise Bonzon (1, 2) , Jérôme Delobelle (3) , Sébastien Konieczny (3) , Nicolas Maudet (4)

1 LIPADE - EA 2517 - Laboratoire d'Informatique Paris Descartes
2 UPD5 - Université Paris Descartes - Paris 5
3 CRIL - Centre de Recherche en Informatique de Lens
4 SMA - Systèmes Multi-Agents

Two kinds of semantics exist for abstract argumentation. Extension-based semantics evaluate the acceptability of sets of arguments, while ranking-based semantics evaluate the strength of each argument. They focus on different aspects of the information conveyed by argumentation systems. After discussing pros and cons of both approaches, we study how to combine them, in order to take benefits from both. We propose six new families of semantics for abstract argumen-tation combining extension-based and ranking-based semantics. More precisely we propose to refine the ranking-based semantics using information coming from extension-based semantics acceptability of arguments, and to modify the extensions chosen by extension-based semantics using preferential information coming from ranking-based semantics.

Domaines: Intelligence artificielle [cs.AI]

Elise Bonzon, Jérôme Delobelle, Sébastien Konieczny, Nicolas Maudet. Combining Extension-based Semantics and Ranking-based Semantics for Abstract Argumentation. 16th International Conference on Principles of Knowledge Representation and Reasoning, Oct 2018, Tempe, United States. ⟨hal-01900735⟩