The Individual and Collective Reasoning Group (ICR) is an interdisciplinary research team at the University of Luxembourg which is driven by the insight that intelligent systems (like humans) are characterized not only by their individual reasoning capacity, but also by their social interaction potential. Its overarching goal is to develop and investigate comprehensive formal models and computational realizations of individual and collective reasoning and rationality.

ICR is anchored in the Lab for Intelligent and Adaptive Systems (ILIAS) of the Department of Computer Science (DCS), and involved in the Interdisciplinary Centre for Security, Reliability and Trust (SnT). The group, which is led by Leon van der Torre, currently counts more than 15 researchers and is strongly engaged in international cooperation.

Our areas are normative multi-agent systems, autonomous cognitive agents, computational social choice, and the foundations of logic-based knowledge representation and reasoning.

Upcoming ICR Events

  • Martin Caminada (ICR Seminar)
    01.08.2022 - 4 p.m.

    Title: The connection between Assumption-Based Argumentation (ABA) and Abstract Argumentation (AA) has been examined in various studies in the literature. In particular, it has been observed that preferred (resp. grounded and stable) semantics for ABA coincides with preferred (resp. grounded and stable) semantics for AA. In the current talk, we will have a closer look at why these semantics coincide. We will make the case that native ABA semantics are essentially defined using a hyperframework (i.e. in terms of a directed hypergraph) using the notion of collective attack. This hyperframework can be translated to a standard argumentation framework by applying the notion of attack semantics. This translation process works fundamentally different from what is usually referred to as the "flattening" of an argumentation framework, as the arrows of the hyperframework become the nodes of the argumentation framework (i.e. no meta arguments in the traditional sense of the word are generated). In the current talk, we examine how this process works and how it can explain the equivalence between preferred (resp. grounded and stable) semantics of ABA and preferred (resp. grounded and stable) semantics of AA. The same process can also be used to explain the non-equivalence between semi-stable semantics for ABA and semi-stable semantics for AA.

    You can join live or via WebEx.