This work discusses the formulation of argumentative dialogue as Markov game. We show how formal systems for persuasive dialogues that adhere to a certain structure can be reformulated as Markov games and thus be addressed as Reinforcement Learning task in a multi-agent setting. We validate our approach on an implementation of a proof of principle scenario where we show that the optimal policy can be learned.
Josh Murphy, Alexandru Burdusel, Michael Luck, Steffen Zschaler, Elizabeth Black
221 - 232
We consider a one-to-many persuasion setting, where a persuader presents arguments to a multi-party audience, aiming to convince them of some particular goal argument. The individual audience members each have differing personal knowledge, which they use, together with the arguments presented by the persuader, to determine whether they are convinced of the goal. The persuader must, therefore, carefully consider its strategy, i.e., which arguments to assert, in order to maximise the number of convinced audience members. Here, we use evolutionary search to find (near-)optimal strategies for the persuader. We implement our approach using search-based model engineering, which provides a natural and efficient encoding for such problems. We investigate the performance of our approach on a range of settings, considering different structures and sizes of argumentation frameworks (representing the underlying knowledge available to the persuader and audience members), and varying the size of audience and of the audience members' personal knowledge bases. We show that we can find effective strategies for problems with more than 200 arguments and more than 100 audience members. Further, we show that the approach supports multiple persuader objectives, finding persuader strategies that aim to minimise arguments to assert while still maximising the number of convinced audience members.
Proposals for strategies for dialogical argumentation often focus on situations where one of the agents wins the dialogue and the other agent loses. Yet in real-world argumentation, it is common for agents to not involve such zero-sum situations. Rather, the agents may enter into a dialogue with divergent but not necessarily opposing views on what is important in the outcomes from the argumentation. In order to model this kind of situation, we investigate a decision-theoretic approach that allows different participants to have different utility evaluations of a dialogue, and for the proponent to model the opponent's utility evaluation in order to optimize the choice of move in the dialogue.
Exploration and information identification constitute challenging research problems, with important applications in sensemaking over structured argumentative dialogues. In this paper, we present the implementation ArgQL, a high-level declarative query language, designed for querying dialogical data, structured in the principles of argumentation. We implement the language using an AIF-based representation and a translation of ArgQL into (complex) SPARQL queries. ArgQL provides a simple and intuitive way to query a structured dialogue using pure argumentative terminology.
Emmanuel Hadoux, Aurélie Beynier, Nicolas Maudet, Paul Weng
249 - 256
Mediation is a process for resolving conflicts among several entities. In argumentation debates, conflicting agents that may be organized as teams exchange arguments to persuade each other. In this paper, we consider an automated mediator, which assigns the speaking slots to agents so as to optimize some objectives and ensure the fairness of the debate. We propose a general setting where the argumentation strategies of the agents are probabilistically known and may evolve over time. We show that the problem can be solved as a semi-Markov decision problem with hidden modes.
Christian Meter, Alexander Schneider, Martin Mauve
257 - 268
Enabling the reuse of arguments as entities that can be shared across multiple Internet-based discussion platforms and that can be improved upon while they are being used and reused has many benefits ranging from easier participation in an online discussion to increasing the quality of arguments. In this paper we propose a mechanism that is able to support the large-scale reuse of arguments by providing distributed version control of argument data. Building on that mechanism we have designed and implemented EDEN, a framework which enables platform providers to easily network their discussions. EDEN is designed for real-world use and provides all tools necessary to enable the reuse of arguments and their interrelation for users and providers alike.
Federico Cerutti, Timothy J. Norman, Alice Toniolo, Stuart E. Middleton
269 - 280
We introduce CISpaces.org, a tool to support situational understanding in intelligence analysis that complements but not replaces human expertise. The system combines natural language processing, argumentation-based reasoning, and natural language generation to produce intelligence reports from social media data, and to record the process of forming hypotheses from relationships among information. In this paper, we show how CISpaces.org meets the desirable requirements elicited from senior professionals, and demonstrate its usage and capabilities to support analysts in delivering effective and tailored intelligence to decision makers.
This paper presents a tool for constructing and evaluating deductive mathematical proofs using formal argumentation called CLEAR (Constructing and evaLuating dEductive mAthematical pRoofs). This tool has a twofold objective: (i) allows students to construct deductive proofs collaboratively using a structured argumentative debate; and (ii) helps instructors to evaluate these proofs and all intermediary steps in order to provide constructive feedbacks to students. This paper focuses on objective (i) and presents results of an experimental study conducted with undergraduate students. The behavior of students during the construction of deductive proofs is analyzed to show whether formal argumentation frameworks allow students to build deductive proofs and measure students' acceptance of CLEAR.
This contribution outlines the diversity of warrants which can connect a claim and a justification. Opposite justifications can support a claim and vice versa: two opposite claims can be supported by the same justification. Since warrants are seldom explicit, this contribution introduces a conceptual model, based on telicity, to automatically induce the content of warrants. This model is evaluated on a corpus of user generated content on various topics. The grammatical and conceptual subtleties of warrants, depending on the argument they support, is also analyzed.
We consider a long-established approach to text analysis applying it to the specific structure of oratory and spoken arguments. This method consists of deriving measures of the so-called “Fractal dimension”. Building on previous work linking “aesthetic appeal” to fractal dimensions of a particular value from the fields of music and literary study, we present empirical analyses of a number of different “persuasive texts” against a range of different choices for determining fractality. Initial results suggest that distinctive “oratorical aims” as may be represented in a text become evident via distinctive fractal dimension.
Jacky Visser, John Lawrence, Jean Wagemans, Chris Reed
313 - 324
In this paper, we present an in-depth comparative analysis of two classifications of argument schemes: Walton's typology and Wagemans' Periodic Table of Arguments. We describe annotation guidelines for each classification and apply these to a corpus of arguments from the 2016 US presidential debates. In so doing, we achieve substantial inter-annotator agreement, and produce what, to the best of our knowledge, are the two largest and most reliably annotated corpora of argument schemes in dialogical argumentation publicly available. In describing the creation and comparison of these corpora, we discuss the strengths of each, with an eye towards both computational modelling and argument mining.
Tobias Krauthoff, Christian Meter, Michael Baurmann, Gregor Betz, Martin Mauve
325 - 336
In this paper, we present D-BAS, a dialog-based online argumentation system, tailored to support e-participation processes. The main idea of D-BAS is to let users exchange proposals and 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 needed to be addressed in order to realize such a system, provide solutions for those challenges, report on a full scale implementation of D-BAS and summarize the findings from a real world e-participation process, where D-BAS provided the infrastructure for online argumentation.
Mark Snaith, Dominic De Franco, Tessa Beinema, Harm op den Akker, Alison Pease
337 - 344
Goal-setting is a frequently adopted strategy in behaviour change coaching. When setting a goal, it is important that it is understood and agreed upon by all parties, and not simply accepted as-is. We present here a dialogue game for multi-party goal-setting, in which multiple health coaches can contribute in order to find a goal that is acceptable to both the patient, and the coaches themselves. Our proposed game incorporates three important aspects of goal-setting and health coaching, (1) coaches can query each other's proposed goals, (2) the patient takes ownership of the goal, and (3) the patient themselves can propose goals.
Alison R. Panisson, Simon Parsons, Peter McBurney, Rafael H. Bordini
345 - 352
Recently, argumentation frameworks have been extended in order to consider trust when defining preferences between arguments, given that arguments (or information that supports the arguments) from more trustworthy sources may be preferred to arguments from less trustworthy sources. Although such literature presents interesting results on argumentation-based reasoning and how agents define preferences between arguments, there is little work taking into account agent strategies for argumentation-based dialogues using such information. In this work, we propose an argumentation framework in which agents consider how much the recipient of an argument trusts others in order to choose the most suitable argument for that particular recipient, i.e., arguments constructed using information from those sources that the recipient trusts. Our approach aims to allow agents to construct more effective arguments, depending on the recipients and on their views on the trustworthiness of potential sources.
In argumentation with incomplete information, an agent often needs to outsource justification of its arguments to other agents, having not sufficient arguments of its own to defend them. Formal characterisation of the impact of such common practice on agents' decision over which of its arguments are acceptable has not been well-investigated. We present an epistemic agent argumentation theory in which an agent can outsource justification of its arguments to its benefactors either by argumentation sharing or relegation. Semantics will be formulated.
Alison R. Panisson, Asad Ali, Peter McBurney, Rafael H. Bordini
361 - 368
One of the main challenges in integrating different smart applications is security. Among the problems related to security, data access control is one of the most important, given that it involves end-users' privacy. In this paper, we propose an approach to the modelling of data access control interfaces using argumentation-based agents. In particular, we introduce argumentation schemes for data access control, which are based on some of the most relevant models for data access control currently available. Our approach considers not only the usual access control policies, describing which category of agents has access to which category of information, but also emergency policies, describing situations where special emergency access control rules apply. Using argumentation-based agents as access control interfaces allows us to deal with the uncertainty about external information, allowing a correct categorisation of agents that request access to information, as well as allowing agents to expose emergency situations in which emergency access control rules may apply.
We propose a new graded semantics for abstract argumentation frameworks that is based on the constellations approach to probabilistic argumentation. Given an abstract argumentation framework, our approach assigns uniform probability to all arguments and then ranks arguments according to the probability of acceptance wrt. some classical semantics. Albeit relying on a simple idea this approach (1) is based on the solid theoretical foundations of probability theory, and (2) complies with many rationality postulates proposed for graded semantics. We also investigate an application of our approach for inconsistency measurement in argumentation frameworks and show that the measure induced by the probabilistic graded semantics also complies with the basic rationality postulates from that area.
Bruno Yun, Srdjan Vesic, Madalina Croitoru, Pierre Bisquert
381 - 392
To address the needs of the EU NoAW project, in this paper we introduce a new modular framework that generates viewpoints (i.e. extensions) based on ranking argumentation semantics by considering a selection function, a ranking on arguments and a lifting function as its input parameters. We study the different combinations of the input parameters and introduce a set of postulates investigated for the framework's different classes of output.
We study rankings over labelings as a generalization of traditional labeling-based semantics in abstract argumentation. Our approach is an alternative to recent developments on rankings over arguments. The formal basis is a qualitative abstraction of probability theory called ranking theory. We propose a fundamental property, called SCC stratification, that a ranking-theoretic semantics can be expected to satisfy, present a general scheme to define a ranking-theoretic semantics, and determine conditions under which this scheme satisfies SCC stratification.
Souhila Kaci, Leendert van der Torre, Serena Villata
405 - 412
Consider an argument A that is attacked by an argument B, while A is preferred to B. Existing approaches will either ignore the attack or reverse it. In this paper we introduce a new reduction of preference and attack to defeat, based on the idea that in such a case, instead of ignoring the attack, the preference is ignored. We compare this new reduction with the two existing ones using a principle-based approach, for the four Dung semantics. The principle-based or axiomatic approach is a methodology to choose an argumentation semantics for a particular application, and to guide the search for new argumentation semantics. For this analysis, we also introduce a fourth reduction, and a semantics for preference-based argumentation based on extension selection. Our classification of twenty alternatives for preference-based abstract argumentation semantics using six principles suggests that our new reduction has some advantages over the existing ones, in the sense that if the set of preferences increases, the sets of accepted arguments increase as well.
In abstract argumentation theory, multiple argumentation semantics have been proposed that allow to select sets of jointly acceptable arguments from a given set of arguments based on the attack relation between arguments. The existence of multiple argumentation semantics raises the question which of these semantics predicts best how humans evaluate arguments, possibly depending on the thematic context of the arguments. In this study we report on an empirical cognitive study in which we tested how humans evaluate sets of arguments depending on the abstract structure of the attack relation between them. Two pilot studies were performed to validate the intended link between argumentation frameworks and sets of natural language arguments. The main experiment involved a group deliberation phase and made use of three different thematic contexts of the argument sets involved. The data strongly suggest that independently of the thematic contexts that we have considered, strong acceptance and strong rejection according to the CF2 and preferred semantics are a better predictor for human argument acceptance than the grounded semantics (which is identical to strong acceptance/rejection with respect to complete semantics). Furthermore, the data suggest that CF2 semantics predicts human argument acceptance better than preferred semantics, but the data for this comparison is limited to a single thematic context.
When confronted with the same abstract argumentation framework, specifying a set of arguments and an attack-relation between them, different agents may disagree on which arguments to accept, i.e., they may choose different extensions. In the context of designing systems to support collective argumentation, we may then wish to aggregate such alternative extensions into a single extension that appropriately reflects the views of the group as a whole. Focusing on a conceptually and computationally simple family of aggregation rules, the quota rules, we analyse under what circumstances relevant properties of extensions shared by all extensions reported by the individual agents will be preserved under aggregation. The properties we consider are the classical properties of argumentation semantics, such as being a conflict-free, a complete, or a preferred extension. We show that, while for some properties there are quota rules that guarantee their preservation, for the more demanding properties it is impossible to do so in general.
Jérémie Dauphin, Marcos Cramer, Leendert Van Der Torre
437 - 444
Most argumentation semantics allow for multiple extensions, which raises the question of how to choose among extensions. We propose to study this question as a decision problem. Inspired by decision trees commonly used in economics, we introduce the notion of a decision graph for deciding between the multiple extensions of a given AF in a given semantics. We distinguish between abstract decision graphs and concrete instantiations thereof. Inspired by the principle-based approach to argumentation, we formulate two principles that mappings from argumentation frameworks to decision graphs should satisfy, the principles of decision-graph directionality and that of directional decision-making. We then propose a concrete instantiation of decision graphs, which satisfies one of these principles. Finally, we discuss the potential for further research based on this novel methodology.
Pietro Baroni, Serena Borsato, Antonio Rago, Francesca Toni
447 - 448
This demo presents the web system “Games of Argumentation”, which allows users to build argumentation graphs and examine them in a game-theoretical manner using up to three different evaluation techniques. The concurrent evaluations of arguments using different techniques, which may be qualitative or quantitative, provides a significant aid to users in both understanding game-theoretical argumentation semantics and pinpointing their differences from alternative semantics, traditional or otherwise, to differentiate between them.