Transparent Rational Decisions by Argumentation

There is a well-documented indication that several applications would benefit from the transparency afforded by argumentation to support decision-making where standard decision theory is not useful, e.g. in healthcare. However, to date, research on argumentation-based decision making has only been partially successful in realising its promise. In our view this is predominantly due to its lack of theoretical validation in the form of rationality properties and its disregard for the interplay between individual rationality and social good when used in collaborative settings. We discuss how to address these challenges for the promise of argumentation-based decision-making to be fully realised as a principled mechanism for transparent and rational decision-making. Context Argumentation, initially studied in philosophy and law [20, 16], has been researched extensively in AI and Computer Science in the last two decades [2, 18]. Simply stated, argumentation focuses on interactions where parties plead for and against some conclusion. In its most abstract form [6], an argumentation framework consists simply of a set of arguments and a binary relation representing the attacks between the arguments, and corresponds to a directed graph (args, attacks), where args is the set of arguments and attacks ⊆ args × args. By instantiating the notion of arguments and the attack relations, different ‘structured’ argumentation frameworks can be obtained, predominantly based on logic, within which arguments have a concrete structure and attacks are determined on the basis of this structure. For example, in assumptionbased argumentation [7], arguments are obtained recursively from a given set of logical rules and are supported by these rules and assumptions. Also, an argument attacks another if the former supports a claim conflicting with some assumption in the latter,