This paper provides a game-theoretical investigation on how to determine optimal strategies in dialogue games for argumentation. To make our ideas as widely applicable possible, we adopt an abstract dialectical setting and model dialogues extensive with perfect information where are determined by preferences over outcomes of the disputes. In turn, specified terms expected utility combining probability success arguments costs benefits associated arguments.

Argumentation is modelled as a game where the payoffs are measured in terms of probability that claimed conclusion is, or not, defeasibly provable, given history arguments have actually been exchanged, and factual premises. The calculated using standard variant Defeasible Logic, combination with calculus. It new element present approach exchange analysed theoretical tools, yielding prescriptive to some extent even predictive account actual course play. A brief comparison existing...

This paper proposes some variants of Temporal Defeasible Logic (TDL) to reason about normative modifications. These make it possible differentiate cases in which, for example, modifications at time change legal rules but their conclusions persist afterwards from where also are blocked.

This paper introduces a class of pseudo-orbits which guarantees the same lower bound error (LBE) for two different natural interval extensions discrete maps. In previous work, LBE was investigated along with simple technique to evaluate numerical accuracy free-run simulations polynomial NARMAX or similar Here we prove that it is possible calculate pseudo-orbits, extending so results work in valid only one pseudo-orbits. The main application this provide estimation LBE. We illustrate our...

Probabilistic argumentation combines the quantitative uncertainty accounted by probability theory with qualitative captured argumentation. In this paper, we investigate problem of learning structure an argumentative graph to account for (a distribution of) labellings a set arguments. We consider general abstract framework, where arguments is left unspecified, and focus on grounded semantics. present, experimental insights, anytime algorithm evaluating `on fly' hypothetical attacks from...

In computational models of argumentation, the justification statements has drawn less attention than construction and arguments. Significant losses sensitivity or expressibility on statement statuses can be incurred by otherwise appealing formalisms. order to reappraise and, more generally, support a uniform modelling different phases argumentation process we introduce multi-labelling systems, generic formalism devoted represent reasoning processes consisting sequence labelling stages....

The outcome of a legal dispute, namely, the decision its adjudicator, is uncertain, and both parties develop their strategies on basis appreciation probability that adjudicator will accept arguments or adversary. Costs gains have to be balanced in light this uncertainty order identify most convenient strategies. This paper provides probabilistic approach embedded into an argumentation framework capture use determine expected utility engage dispute.

Probabilistic argumentation and neuro-argumentative systems offer new computational perspectives for the theory applications of argumentation, but their principled construction involves two entangled problems.On one hand, probabilistic aims at combining quantitative uncertainty addressed by probability with qualitative dependences amongst arguments as well learning are usually neglected.On other opportunity to couple advantages massive parallel computation from neural networks argumentative...

Towards neuro-argumentative agents based on the seamless integration of neural networks and defeasible formalisms, with principled probabilistic settings along efficient algorithms, we investigate argumentative Boltzmann machines where possible states a machine are constrained by prior knowledge. To make our ideas as widely applicable possible, acknowledging role sub-arguments in argumentation, consider an abstract argumentation framework accounting for sub-arguments, but content...

Reinforcement learning is a widespread mechanism for adapting the individual behaviour of autonomous agents, while norms are well-established means organising common conduct these agents. Therefore, norm-governed reinforcement agents appear to be powerful bio-inspired, as well socio-inspired, paradigm construction decentralised, self-adapting, self-organising systems. However, convergence and not straightforward it appears: can 'misguide' development norms, 'stall' optimal behaviour. In this...