Dominant Decisions by Argumentation Agents

Abstract

Abstract. We introduce a special family of (assumption-based argumentation) frameworks for reasoning about the bene ts of decisions. These frameworks can be used for representing the knowledge of intelligent agents that can autonomously choose the \best" decisions, given subjective needs and preferences of decision-makers they \represent". We understand \best" decisions as dominant ones, giving more bene ts than any other decisions. Dominant decisions correspond, within the family of argumentation frameworks considered, to admissible arguments. We also propose the use of degrees of admissibility of arguments as a heuristic to assess subjectively the value of decisions and rank them from \best" (dominant) to \worst". We extend this method to provide notion of relative value of decisions where preferences over bene ts are taken into account. Finally, we show how our techniques can be successfully applied to the problem of selecting satellite images to monitor oil spills, to support electronic marketplaces for earth observation products.1 Introduction This paper presents two methods for evaluating and ranking (respectively) decisions using assumption-based argumentation (ABA) frameworks [4,13,14]. These ABA frameworks can be used for representing the knowledge of intelligent agents that autonomously make the best decision, e.g. by choosing the \best" items/ services available in a service-oriented architecture, e.g. as envisaged in [32]. We de ne a notion of dominance to characterise \best" items, and show that these items are those that correspond, within the family of argumentation frameworks considered, to admissible arguments as de ned in [4,13]. Intuitively, the method relies upon comparing the value of di erent items by \arguing" about the bene ts they provide. The reason why an item provides a bene t is explained logically (by means of arguments). ABA allows to study collections of logically constructed arguments based upon assumptions. As we shall see, ABA frameworks are particularly adequate for representing and reasoning about the bene ts of items as well as for providing explanations to users. In practice, decision-makers do not always choose top-quality (\best") items, but often choose items which t their budget and have highest quality/price ratio. In this context, it is useful to assess numerically the quality or relative quality of items. We propose a novel criterion (or \semantics", as understood in argumentation) for choosing sets of assumptions and arguments. This criterion

A tutorial on assumption-based argumentation

“…All the semantics considered earlier, in terms of sets of arguments or sets of assumptions, can be equivalently reformulated in terms of backward-style and forward-style arguments, by virtue of the correspondence results between the various notions of argument given earlier in Section 4. give examples of use of ABA to model reasoning with argument scheme, decision making, dispute resolution and game-theoretic notions. Furthermore, Matt, Toni, and Vaccari (2010), Matt, Toni, Stournaras, and Dimitrelos (2008), , Fan, Craven, Singer, Toni, and Williams (2013) give several uses of ABA to model decision-making, for various notions of dominant decisions. This section complements these works by illustrating the use of ABA for default reasoning, defeasible reasoning and persuasion.…”

A probabilistic approach to modelling uncertain logical arguments

“…For this, we could investigate extending the approach of decision theoretic argumentation of Amgoud and Prade [9] with implicit and explicit uncertainty, with the problem of generating intentions in the context of uncertainty [7], and/or the optimization of decision making using assumption-based argumentation of Matt et al [42].…”

An explainable multi-attribute decision model based on argumentation

“…we rank decisions on the basis of decision-making criteria with natural language explanations for pairwise comparison provided. Matt et al (2009) introduce a special family of ABA frameworks for reasoning about the benefits of decisions, and thus view, as we do, best decisions as dominant decisions corresponding, within the family of argumentation frameworks considered, to admissible arguments. However, our best decisions are not only dominant, but also minimally redundant.…”

“…Argumentation has been seen as an effective means to facilitate many aspects of decision-making and decision-support systems (Fox et al, 2010;Nawwab et al, 2008), especially when decisions recommended by such systems need to be explained. However, with few exceptions (notably Amgoud and Prade (2009) ;Matt et al (2009) ;; Visser et al (2013)), it is unclear whether or not the output decisions of argumentation-based methods can be deemed to be rational in some decisiontheoretic sense. Furthermore, current works in argumentation-based decision-making either fail to automatically generate explanations (e.g.…”