Program
The times below are in Central European Time (CET).09:30-09.40: opening
09.40-10.20: Ken Satoh, Arranging Issues in Civil Litigation by PROLEG (abstract)(Slides)
10.20-11.00: Huimin Dong, Logic of Defeasible Permission and its Dynamics (abstract)(Slides)
11.00-11.10: Break 11.10-11.50: Emil Weydert, A Lesson in Default Reasoning - What to agree upon (abstract)(Slides)
11.50-12.30: Shohreh Haddadan, Argument Mining and Challenges from Political Debates (abstract)(Slides)
12.30-14.00: Lunch break
14.00-14.40: Marcos Cramer, Comparing the Cognitive Plausibility of Abstract Argumentation Semantics Based on Empirical Studies (abstract)(Slides)
14.40-15.20: Jeremie Dauphin, An Enriched Argumentation Framework with Higher-Level Relations (abstract)(Slides)
15.20-15.30: Break 15.30-16.10: Egberdien van der Torre, The Peripeteia in HCI (abstract)(Slides)
16.10-16.50: Dov Gabbay, Critical Expansions Methodology at the Service of Formal Argumentation (abstract)(Slides)
16.50-17.00: closing
Abstracts:
Prof. Ken Satoh
National Institute of Informatics, Tokyo, JapanArranging Issues in Civil Litigation by PROLEG
We propose a system that enables online argumentation between plaintiff and defendant instead of presenting their claims in court face to face. This system uses a graphic editor to repeat assertion defenses against each other to gradually create a whole argumentation. It can automatically indicate required facts to be proved and possible counter-arguments with its target issue, and clarify current dispute issues. This can help to make a plan for arranging issues in Japanese legal system for judges, to educate law school students and possibly for lay people to make an appropriate argument in litigation.
Dr. Huimin Dong
Zhejiang University, Hangzhou, ChinaLogic of Defeasible Permission and its Dynamics
In this talk, I will develop a deontic logic for defeasible permission and study changes of norms in various updated semantics. When one admits that π or π is permitted, normally, it goes together with the conjunctions of a permission of π and that of π. In the monotonic reasoning on this permission, a permission of π leads to a permission of π and π; while a prohibition of π is introduced, we get into trouble. This is the paradox of free choice permission. Many solutions have been proposed, but a systematic account of handling defeasible permission with norm change is still needed. This paper first introduces the notion of normality to develop a sound and complete deontic logic for defeasible permission, which can be used to analyze some notions in natural language and in games. Further, following Lewisβ idea of changing norms, a systematic way is proposed to capture various dynamics for updating permission and obligation.
Dr. Emil Weydert
University of Luxembourg, Esch-sur-Alzette, LuxembourgA Lesson in Default Reasoning - What to agree upon
Some old facts and recent insights about defaults and reasoning with and about them, with a focus on semantic approaches exploiting ranking measures. In particular, we explore the relation between the monotonic logics of defaults and nonmonotonic reasoning with defaults.
Shohreh Haddadan
University of Luxembourg, Esch-sur-Alzette, LuxembourgArgument Mining and Challenges from Political Debates
Argument mining has been trending recently as a sub-field of natural language processing. Extracting arguments depending on the domain can be challenging from different perspectives. One of these challenges is how to represent the argument structures which are the ultimate output of an argument mining pipeline. In this talk I will present a few examples of such methods. I also talk about the inherent challenges which an argument mining algorithm encounter while dealing with data as practical and concrete as transcripts of political debates data. At the end, I will talk about a few applications which may benefit from automatic argument mining from political debates data.
Dr. Marcos Cramer
TU Dresden, Dresden, GermanyComparing the Cognitive Plausibility of Abstract Argumentation Semantics Based on Empirical Studies
Which argumentation semantics corresponds best to how humans judge the acceptability of arguments? This research question has been addressed by multiple cognitive empirical studies. In this talk I will give an overview over the findings of these studies and make an attempt at synthesizing their results into a coherent picture. One important lesson is that naive-based semantics like CF2 and SCF2 seem to correspond better to human judgements than admissibility-based semantics like preferred and semi-stable. I will also address the question of how future studies on this topic should be designed in order to give us a more complete picture of this issue.
Jeremie Dauphin
University of Luxembourg, Esch-sur-Alzette, LuxembourgAn Enriched Argumentation Framework with Higher-Level Relations
Multiple extensions of Dung's argumentation frameworks (AFs) have been proposed in order to model features of argumentation that cannot be directly modeled in AFs. In order to faithfully model the interaction between explanation and argumentation in scientific debates, AFs have been extended to Explanatory Argumentation Frameworks (EAFs) by adding an explanation relation and an incompatibility relation. Other extensions of AFs studied in the literature include higher-order (recursive) attacks, joint attacks (i.e. attacks from a set of arguments), disjunctive attacks (i.e. attacks on a set of arguments), deductive support and necessary support. In this talk, we generalize these enrichments by also defining higher-order set relations of support, explanation and incompatibility. We combine all these features in a single kind of framework called Extended Explanatory Argumentation Frameworks (EEAFs). We define a labelling semantics for the enriched framework, then additionally make use of the meta-argumentation methodology of flattening, and show that a correspondence between both approaches.
Egberdien van der Torre
Conceptual artist for the Esch 2022 pavilion project University of Luxembourg, Esch-sur-Alzette, LuxembourgThe Peripeteia in HCI
In the field of human-computer interaction, a Wizard of Oz experiment allows researchers to test a concept by having a hidden human programmer and a user who thinks that he is interacting with the machine itself. In this talk we focus on how the communicative content of the interaction can evolve from normal to absurd.
Prof. Dov Gabbay
King's College, London, United Kingdom Bar-Ilan University, Ramat Gan, IsraelCritical Expansions Methodology at the Service of Formal Argumentation
The second half of the previous century has seen a meteoric rise in the landscape of logical systems trying to model applications in computer science AI, language analysis, KR, formal philosophy and more. Fortunately, some methodological order and evaluation could be introduced using Classical Logic and its proof theories.
In this century there was the rise of many formal argumentation systems and semantics addressing a variety of applications, as well as absorbing many aspects of the new logics of the previous century.
There is now the need for a methodology to play the role/provide a home for comparison and evaluation, a role similar to that of classical logic in the previous century.
We offer the methodology of critical expansions:
A generic enhanced argumentation network can be presented schematically in the form (S, R, Enhancements),
Where S is the set of arguments arising/mirroring from some intended application area, R is the basic binary attack relation on S and Enhancements is the various additional aspects coming from the application.
A critical expansion for (S, R, Enhancements) has the form (S, S*, R*) , where S* enlarges S with additional service arguments, and where R* is the attack relation on S*, devised in such a way that (S*, R*) is a critical faithful expansion of (S, R, Enhancements) representing of the enhancements of (S, R, Enhancements) by means of the service arguments. So the enhancements are implemented by the way S* interacts with S using R*.
The lecture will illustrate this idea (using examples) for higher level attacks (any level) and for resolution of cycles.