Weighted argumentation

When dealing with Abstract Argumentation, having preference values on arguments/attacks clearly brings more information to a framework, which can be considered as a directed graph. One of the advantages is the possibility to define a different notion of defence, checking also if the associated preference is stronger than the preference of the considered attack. In the real-world, such values can be represented by “likes” in social-networks, or generic votes in favour of attacks. We focus on qualitative/quantitative preference values on attacks, which indicate their (relative) strength and can measure an argument-pair inconsistency degree. Once assembled, also by moving values from arguments to attacks, it is then possible to redefine semantics, relax the notion of weighted acceptability, and check well-known properties as in Dung’s frameworks, e.g., if a framework is well-founded.