Ebook: Computational Models of Argument

The investigation of computational models of argument is a rich, interdisciplinary, and fascinating research field whose ultimate aim is to support the development of computer-based systems able to engage argumentation-related activities with human users or among themselves. This ambitious research goal involves the study of natural, artificial, and theoretical argumentation and, as such, requires openness to interactions with a variety of disciplines ranging from philosophy and cognitive science to formal logic and graph theory, to mention some.

The biennial International Conference on Computational Models of Argument (COMMA), reaching its sixth edition, provides since ten years a dedicated forum for presentation of the latest advancements in this multifaceted field, covering both basic research and innovative applications.

The first COMMA was supported by the EU 6th Framework Programme project ASPIC and was hosted by the University of Liverpool in 2006 with a vision for the future. After the event, a steering committee promoting the continuation of the conference was established and, since then, the steady growth of interest in computational argumentation research worldwide has gone hand in hand with the development of the conference itself and of related activities by its community.

Since the second edition, organized by IRIT in Toulouse in 2008, plenary invited talks by world-leading researchers and a software demonstration session became an integral part of the conference program.

The third edition, organized in 2010 by the University of Brescia in Desenzano del Garda, saw the addition of a best student paper award. The same year, the new journal Argument and Computation, closely related to the COMMA community, was started.

Since the fourth edition, organized by the Vienna University of Technology in 2012, an Innovative Application Track and a section for Demonstration Abstracts were included in the proceedings.

At the fifth edition, co-organized in 2014 by the Universities of Aberdeen and Dundee in Pitlochry, the main conference was preceded by the first Summer School on Argumentation: Computational and Linguistic Perspectives. The same year, the first International Competition on Computational Models of Argumentation, to be held in 2015, was launched.

This year COMMA is hosted by the University of Potsdam and the conference program is complemented by two satellite workshops, in addition to the second edition of the summer school. Moreover, reflecting the evolution of research publishing worldwide, COMMA 2016 proceedings will be Open Access.

The evolution of COMMA into an articulated event, however, is only subsidiary to the fulfillment of its mission, namely documenting and stimulating the advancement of knowledge and the development of applications in the field.

The past conference programs, along with the present one, give comfortable indications in this respect.

First of all, they have seen, since the very first edition, a balanced blend of theoretical and application-oriented works.

Further, in addition to “traditional” investigation topics in the field, like abstract argumentation frameworks, the conference has always included contributions concerning emerging trends and the development of new connections with other areas.

Among them, it is possible to mention the investigation of a variety of quantitative approaches to argumentation, in relationship with Bayesian networks, probability theory, or fuzzy logic. Also, we wish to mention the area of argument mining, the automatic detection and analysis of argumentation in linguistic data. While the term was barely known three years ago, this research has emerged as a fast-growing subfield of Computational Linguistics (CL), with a variety of specialized workshops having been formed over the past few years, and the topic has also been established in important CL conferences. So far, following the general trend in CL, the methods being applied to argument mining largely rely on machine learning over surface-oriented features of text; but there seems to be great potential in linking the text analysis also to the “deeper” phenomena – reasoning and inference – that render an argumentation plausible.

We conclude by remarking that the success of a conference depends on the contributions of many people.

We acknowledge steady support and encouragement by the COMMA Steering Committee.

We would like to thank the invited speakers, Jens Allwood, Anthony Hunter, and Marie-Francine Moens, for accepting our invitation and for witnessing, once again, the rich diversity of this area with their talks, covering respectively an insightful analysis of the normative and descriptive perspectives in argumentation studies, the promise and challenge of using computational persuasion for applications in behaviour change, and the formidable question of how can a machine acquire world and common sense knowledge for argument mining.

We are deeply grateful to the members of the Program Committee and to the additional reviewers for their invaluable efforts. Their reports and subsequent discussions led to the selection, out of 63 submissions, of 25 full papers and 17 short papers, to be included in the conference proceedings together with 10 demonstration abstracts. The submission and reviewing process has been managed through the Easychair conference system, which we acknowledge for supporting COMMA since the first edition.

Last but not least, we thank all the authors for contributing to the success of the conference with their hard work and commitment.

Berlin/Brescia/Potsdam, July 2016

Pietro Baroni (Program chair)

Thomas F. Gordon (Conference chair)

Tatjana Scheffler (Local organization co-chair)

Manfred Stede (Conference chair)

Persuasion is an activity that involves one party trying to induce another party to believe something or to do something. It is an important and multifaceted human facility. Obviously, sales and marketing is heavily dependent on persuasion. But many other activities involve persuasion such as a doctor persuading a patient to drink less alcohol, a road safety expert persuading drivers to not text while driving, or an online safety expert persuading users of social media sites to not reveal too much personal information online. As computing becomes involved in every sphere of life, so too is persuasion a target for applying computer-based solutions. An automated persuasion system (APS) is a system that can engage in a dialogue with a user (the persuadee) in order to persuade the persuadee to do (or not do) some action or to believe (or not believe) something. To do this, an APS aims to use convincing arguments in order to persuade the persuadee. Computational persuasion is the study of formal models of dialogues involving arguments and counterarguments, of user models, and strategies, for APSs. A promising application area for computational persuasion is in behaviour change. Within healthcare organizations, government agencies, and non-governmental agencies, there is much interest in changing behaviour of particular groups of people away from actions that are harmful to themselves and/or to others around them.

The problem of understanding the stream of messages exchanged on social media such as Facebook and Twitter is becoming a major challenge for automated systems. The tremendous amount of data exchanged on these platforms as well as the specific form of language adopted by social media users constitute a new challenging context for existing argument mining techniques. In this paper, we describe an ongoing work towards the creation of a complete argument mining pipeline over Twitter messages: (i) we identify which tweets can be considered as arguments and which cannot, (ii) over the set of tweet-arguments, we group them by topic, and (iii) we predict whether such tweets support or attack each other. The final goal is to compute the set of tweets which are widely recognized as accepted, and the different (possibly conflicting) viewpoints that emerge on a topic, given a stream of messages.

In this position paper we propose a novel approach to online argumentation. It avoids the pitfalls of unstructured systems such as asynchronous threaded discussions and it is usable by any participant without training while still supporting the full complexity of real-world argumentation. The key idea is to let users exchange arguments with each other in the form of a time-shifted dialog where arguments are presented and acted upon one-at-a-time. We highlight the key research challenges that need to be addressed in order to realize such a system and provide first solutions for those challenges.

Group polarization occurs when an initial attitude or belief of individuals becomes more radical after group discussion. Polarization often leads subgroups towards opposite directions. Since the 1960s this effect has been observed and repeatedly confirmed in lab experiments by social psychologists. Persuasive Arguments Theory (PAT) emerged as the most convincing explanation for this phenomenon. This paper is a first attempt to frame the PAT explanation more formally by means of Bipolar Argumentation Frameworks (BAFs). In particular, I show that polarization may emerge in a BAF by simple and rational belief updates by participants.

The increase in routine clinical data collection coupled with an expectation to exploit this in support of evidence based decision making creates a need for an intelligent model selection system to support clinicians when analysing data because clinicians often lack the statistical expertise to do this independently. In a previous position paper, an argumentation based approach to devise a decision support system for such an application was introduced. This approach ignored the relative strength of arguments for and against alternative models. This paper demonstrates how an extended argumentation framework can be employed to capture and reason with statistical and research domain knowledge that affects the relative strength of arguments. The approach is validated by means of a real-world case study.

In this paper we demonstrate how to benefit from structured argumentation frameworks and their implementations to provide for reasoning capabilities of Ontology Based Data Access systems under inconsistency tolerant semantics. More precisely, given an inconsistent Datalog^± knowledge base we instantiate it using the ASPIC⁺ framework and show that the reasoning provided by ASPIC⁺ is equivalent to the main inconsistent tolerant semantics in the literature. We provide a workflow that shows the practical interoperability of the logic based frameworks handling Datalog^± and ASPIC⁺.

In practical reasoning, it is important to take into consideration what other agents will do, since this will often influence the effect of actions performed by the agent concerned. In previous treatments, the actions of others must either be assumed, or argued for using a similar form of practical reasoning. Such arguments, however, will also depend on assumptions about the beliefs, values and preferences of the other agents, and so are difficult to justify. In this paper we capture, in the form of argumentation schemes, reasoning about what others will do, which depends not on assuming particular actions, but through consideration of the expected utility (based on the promotion and demotion of values) of particular actions and alternatives. Such arguments depend only on the values and preferences of the agent concerned, and do not require assumptions about the beliefs, values and preferences of the other relevant agents. We illustrate the approach with a running example based on Prisoner's Dilemma.

Dung's abstract argumentation theory is a widely used formalism to model conflicting information and to draw conclusions in such situations. Hereby, the knowledge is represented by argumentation frameworks (AFs) and the reasoning is done via semantics extracting acceptable sets. All reasonable semantics are based on the notion of conflict-freeness which means that arguments are only jointly acceptable when they are not linked within the AF. In this paper, we study the question which information on top of conflict-free sets is needed to compute extensions of a semantics at hand. We introduce a hierarchy of verification classes specifying the required amount of information and show that well-known semantics are exactly verifiable through a certain such class. This also gives a means to study semantics lying between known semantics, thus contributing to a more abstract understanding of the different features argumentation semantics offer.

In this paper, we propose a way to derive constraints for a Bayesian Network from structured arguments. Argumentation and Bayesian networks can both be considered decision support techniques, but are typically used by experts with different backgrounds. Bayesian network experts have the mathematical skills to understand and construct such networks, but lack expertise in the application domain; domain experts may feel more comfortable with argumentation approaches. Our proposed method allows us to check Bayesian networks given arguments constructed for the same problem, and also allows for transforming arguments into a Bayesian network structure, thereby facilitating Bayesian network construction.

Many computational problems in the area of abstract argumentation are intractable. For some semantics like preferred and semi-stable, important decision problems can even be hard for classes of the second level of the polynomial hierarchy. One approach to deal with this inherent difficulty is to exploit structure of argumentation frameworks. In particular, algorithms that run in linear time for argumentation frameworks of bounded treewidth have been proposed for several semantics. In this paper, we contribute to this line of research and propose a novel algorithm for the semi-stable semantics. We also present an implementation of the algorithm and report on some experimental results.

A key area in the research agenda of modelling argumentation is to accurately model argumentation on the social web. In this paper we propose additional extensions to our ontology for argumentation on the social web (which integrates elements of the Argument Interchange Format and the Semantically Interlinked Online Communities project) for the purposes of modelling social and rhetorical tactics used in eristic or irrational arguments. We then present a review of these extensions from a panel of experts in the fields of argumentation modelling, web science, philosophy and open and linked data and discuss the value of modelling social argument, the challenges faced to create usable and accurate models and the completeness, clarity and consistency of our proposed additions.

We establish a uniform modular translation of Abstract Dialectical Frameworks into the formalism of the causal calculus, and discuss the correspondences this translation creates between a number of semantics suggested for ADFs and their causal counterparts.

Argumentation is based on the exchange and the evaluation of interacting arguments. Unlike Dung's theory where arguments are either accepted or rejected, ranking-based semantics rank-order arguments from the most to the least acceptable ones. We propose in this work six new ranking-based semantics. We argue that, contrarily to existing ranking semantics in the literature, that focus on evaluating attacks and defenses only, it is reasonable to give a prominent role to non-attacked arguments, as it is the case in standard Dung's semantics. Our six semantics are based on the propagation of the weight of each argument to its neighbors, where the weight of non-attacked arguments is greater than the attacked ones.

Argumentation has become an important topic in artificial intelligence; the basic idea is to identify arguments in favor and against a statement, select the acceptable ones, and determine whether the original statement can be accepted or not. However, the arguments involved in an argumentative discussion may have different relevance degrees; for this reason, argumentation frameworks need to represent the qualities that describe the soundness of an argument in order to refine the acceptability process performed over the argumentation model.

Cost based methods have been proposed as ways to extend Argumentation frameworks proposed by Dung but have been found to contain an anomaly. In this paper, we introduce a new cost based method that is free from the anomaly. We also describe a preliminary algorithm to calculate the cost.

Spectral analysis – the study of the properties of the eigenvalues associated with some matrix derived from an underlying graph form – has proven to offer valuable insights in many domains where graph-theoretic models are prevalent. Abstract argumentation frameworks (afs) are, of course, one such model and have provided a unifying basis for defining semantic properties related to concepts of “argument acceptability”. In this paper we consider the possible benefits of adopting spectral methods as a tool for analysing argumentation structures, presenting a preliminary empirical study of semantics in afs and properties of the associated spectrum.

The current paper provides a dialectical interpretation of the argumentation-based judgment aggregation operators of Caminada and Pigozzi. In particular, we define discussion-based proof procedures for the foundational concepts of down-admissible and up-complete. We then show how these proof procedures can be used as the basis of dialectical proof procedures for the sceptical, credulous and super credulous judgment aggregation operators.

This paper proposes a new framework able to take into account recursive interactions in bipolar abstract argumentation systems. We address issues such as “How an interaction can impact another one?”, or in other words “How can the validity of an interaction be affected if this interaction is attacked or supported by another one?”. Thus, building on numerous examples, a new method for flattening such recursive bipolar abstract argumentation systems (ASAF) using meta-arguments is proposed and compared with the original framework defined in [8].

In this paper we investigate the impact of automated configuration techniques on the ArgSemSAT solver—runner-up of the ICCMA 2015—for solving the enumeration of preferred extensions. Moreover, we introduce a fully automated method for varying how argumentation frameworks are represented in the input file, and evaluate how the joint configuration of frameworks and ArgSemSAT parameters can have a remarkable impact on performance. Our findings suggest that automated configuration techniques lead to improved performances in argumentation solvers, an important message for participants to the forthcoming competition.