Workshop on New Trends in Formal Argumentation

The 2nd workshop on new trends in formal argumentation will take place in Vienna on April 30, 2019.

ORGANIZATION

Stefan Woltran
Anna Rapberger
Martin Diller

AIMS AND SCOPE

Argumentation is nowadays a central part of modern Artificial Intelligence research, and a main topic at major AI conferences, with application domains in legal reasoning, multi-agent systems, and decision support. Key to this research is the field of formal argumentation which studies formal models to represent arguments, methods to infer conclusions from conflicting points of view, and to provide means for automated reasoning for computation of acceptable conclusions. In the past decades many foundational results have been established for formal argumentation, yet several challenges only have recently found more attention in the research community, or remain not sufficiently solved, such as formalizations of dynamics in argumentation, enrichment and in-depth study of existing formal models to close gaps to applications, efficient algorithms and system implementations, and many more.

The aim of this workshop is to bring together researchers from several parts of the diverse field of formal argumentation in order to discuss recent results and ongoing work on the new challenges.

Topics of interest include, but are not limited to:

Formal models in argumentation
Structured argumentation
Argumentation and multi-agent systems
Preferences in argumentation
Dynamics in argumentation
Applications of argumentation
Argumentation and natural language
Complexity of problems in argumentation
Implementations and argumentation systems

INVITED SPEAKERS

We are happy to announce that the following speakers will give invited talks:

Federico Cerutti, Cardiff University,
Sylwia Polberg , Cardiff University,
Matthias Thimm , Universität Koblenz-Landau.

PRELIMINARY PROGRAM AND SCHEDULE

14:00	Invited talk: Federico Cerutti	Probabilistic Logic Programming with Beta-Distributed Random Variables and Its Connection with Abstract Argumentation
14:40	Invited talk: Matthias Thimm	Approximate Methods for Argumentative Reasoning
15:20	Carlo Taticchi	A Cooperative-game Approach to Share Acceptability and Rank Arguments
15:40	Coffee break
16:05	Invited talk: Sylwia Polberg	Framework for Computational Persuasion
16:45	Atefeh Keshavarzi Zafarghandi	Abstract Dialectical frameworks and discussion games
17:05	Johannes P. Wallner	Reasoning over Assumption-Based Argumentation Frameworks via Direct Answer Set Programming Encodings
17:25	Stefan Woltran	Abstract Argumentation in the Argumentation Pipeline - Mind the Gap!
17:45	Closing

ABSTRACTS OF THE TALKS

Invited Talks

Probabilistic Logic Programming with Beta-Distributed Random Variables and Its Connection with Abstract Argumentation
Federico Cerutti, Cardiff University

Abstract: We enable aProbLog---a probabilistic logical programming approach---to reason in presence of uncertain probabilities represented as Beta-distributed random variables. We achieve the same performance of state-of-the-art algorithms for highly specified and engineered domains, while simultaneously we maintain the flexibility offered by aProbLog in handling complex relational domains. Our motivation is that faithfully capturing the distribution of probabilities is necessary to compute an expected utility for effective decision making under uncertainty: unfortunately, these probability distributions can be highly uncertain due to sparse data. To understand and accurately manipulate such probability distributions we need a well-defined theoretical framework that is provided by the Beta distribution, which specifies a distribution of probabilities representing all the possible values of a probability when the exact value is unknown. Finally, we will discuss implications in particular in the context of the constellation approach to abstract argumentation.

Approximate Methods for Argumentative Reasoning
Matthias Thimm, Universität Koblenz-Landau

Abstract: Reasoning with formal accounts to argumentation is a computationally demanding task. Many problems pertaining to abstract argumentation have been shown to be intractable in general and reasoning with structured accounts to argumentation suffer, in addition, from a possibly exponential number of arguments that have to be considered. Recents advances have shown that complete solvers for abstract argumentation can address challenging problem sizes, but there is still a lot of work to do. In this talk I address approximation techniques to solve such problems, i.e., methods that are not necessarily sound or complete in general, but approximate correct answers as best as possible while being tractable. For that, I will present recent and ongoing works on using stochastic local search as well as graph convolutional neural networks for solving reasoning problems in abstract argumentation. Moreover, I will discuss ongoing works on approximation techniques for structured argumentation approaches such as ASPIC+.

Framework for Computational Persuasion
Sylwia Polberg, Cardiff University

Abstract: The aim of behaviour change is to help people to change aspects of their behaviour for the better (e.g., to decrease calorie intake, to drink in moderation, etc.). In current persuasion technology for behaviour change, the emphasis is on helping people to explore their issues (e.g., through questionnaires or game playing) or to remember to follow a behaviour change plan (e.g., diaries and email reminders). However, recent development in computational modelling of argument are leading to an alternative approach to persuasion that can potentially be harnessed in behaviour change applications. In this approach, a software system and a user can exchange arguments in a dialogue in order to investigate a given issue in a tailored and dynamic manner. By gaining information about the users perspective, the system can provide arguments to fill gaps in the users knowledge, and to overturn misconceptions held by the user. In this talk, I will discuss the concept of computational persuasion and summarize the outcomes of the recently concluded project in that area. The results involve both theoretical and practical advancements as well as series of empirical studies with participants.

Talks

A Cooperative-game Approach to Share Acceptability and Rank Arguments
Carlo Taticchi, Gran Sasso Science Institute

Abstract: We deploy a game-theoretic approach for analysing the acceptability of arguments in a generic Abstract Argumentation Framework. The result is a ranking-based semantics, which sorts arguments from the most to the least acceptable. In the computation of such a ranking, we adopt the Shapley Value formula, since it is usually used to fairly distribute costs to several entities in coalitions (labelled sets of arguments in our case). Finally, we show that some well-known properties are satisfied by the ranked-semantics we designed, and we provide an example of how our approach works.

Joint work with Stefano Bistarelli, Paolo Giuliodori, and Francesco Santini.

Abstract Dialectical frameworks and discussion games
Atefeh Keshavarzi Zafarghandi, University of Groningen

Abstract: Since Dung introduced abstract argumentation frameworks (AFs) as a formal model of argumentation, several generalizations of AFs have been proposed. Abstract dialectical frameworks (ADFs) are a particularly expressive generalization of AFs, which can represent relations among arguments beyond simple attacks.
Basically, the term 'dialectical method' refers to a discussion among two or more people who have different points of view about a subject but are willing to find out the truth by argumentation. Thus, in classical philosophy, dialectic is a method of reasoning based on arguments and counter-arguments. In ADFs, dialectical methods implicitly play a role in determining the truth-value of arguments via evaluation methods encoded in several "semantics".
However, as far as we know, the relation between ADFs and dialectical methods in the sense of discussions among agents has not been made explicit yet. One of the ways of making this connection explicit, which we take up in this work, is characterising the semantics of ADFs in terms of discussion games. To this end, we focus on preferred semantics and we show that preferred semantics of ADFs are indeed expressible in terms of discussion games. That is, we show from the perspective of game theory that there exists a proof strategy for arguments that are credulously accepted (denied) under the preferred semantics. To show that the preferred discussion games are comprehensive enough, we show soundness and completeness of the proof method.

Joint work with Bart Verheij and Rineke Verbrugge.

Reasoning over Assumption-Based Argumentation Frameworks via Direct Answer Set Programming Encodings
Johannes P. Waller, TU Wien

Abstract: Focusing on assumption-based argumentation (ABA) as a central structured formalism to AI argumentation, we propose a new approach to reasoning in ABA with and without preferences. While previous approaches apply either specialized algorithms or translate ABA reasoning to reasoning over abstract argumentation frameworks, we develop a direct approach by encoding ABA reasoning tasks in answer set programming. This significantly improves on the empirical performance of current ABA reasoning systems. We also give new complexity results for reasoning in ABA+, suggesting that the integration of preferential information into ABA results in increased problem complexity for several central argumentation semantics.

Abstract Argumentation in the Argumentation Pipeline - Mind the Gap!
Stefan Woltran, TU Wien

Abstract: Abstract argumentation frameworks have been introduced by Dung as part of an argumentation process, where arguments and conflicts are derived from a given knowledge base. It is solely this relation between arguments that is then used in order to identify acceptable sets of arguments. A final step concerns the acceptance status of particular statements by reviewing the actual contents of the acceptable arguments. However, a certain subtlety - which is often neglected -- occurs when different arguments with the same claim are constructed during this process. On the one hand, it makes the actual notion of acceptance problems ambiguous and we show that a claim-centric definition of acceptance leads to changes compared to the known complexity landscape for abstract argumentation. On the other hand, it is natural to ask under which circumstances such situations can be avoided, i.e. when can the abstract step be done in such a way that each argument represents a unique claim. We show, for all standard semantics, that scenarios where arguments with the same claim have the same outgoing attacks, can be equivalently represented as argumentation frameworks with collective attacks but unique claims.

Joint work with Wolfgang Dvořák and Anna Rapberger.

PARTICIPANTS

Federico Cerutti,
Martin Diller,
Wolfgang Dvořák,
Uwe Egly,
Atefeh Keshavarzi Zafarghandi,
Anna Rapberger,
Sylwia Polberg,
Jörg Pührer,
Carlo Taticchi,
Matthias Thimm,
Johannes Wallner,
Stefan Woltran

VENUE

Seminarraum Menger
Address: Favoritenstraße 9-11, TU Vienna, Vienna (map)
Third floor (map)
Room: HF0311

PHOTOS FROM THE EVENT

Federico Cerutti and Matthias Thimm

The regular speakers

The audience