Gerard A. W. Vreeswijk

Logics for Defeasible Argumentation

Logic is the science that deals with the formal principles and criteria of validity of patterns of inference. This chapter surveys logics for a particular group of patterns of inference, namely those where arguments for and against a certain claim are produced and evaluated, to test the tenability of the claim. Such reasoning processes are usually analysed under the common term ‘defeasible argumentation’. We shall illustrate this form of reasoning with a dispute between two persons, A and B. They disagree on whether it is morally acceptable for a newspaper to publish a certain piece of information concerning a politician’s private life.1 Let us assume that the two parties have reached agreement on the following points.

Towards an argument interchange format

The theory of argumentation is a rich, interdisciplinary area of research straddling the fields of artificial intelligence, philosophy, communication studies, linguistics and psychology. In the last few years, significant progress has been made in understanding the theoretical properties of different argumentation logics. However, one major barrier to the development and practical deployment of argumentation systems is the lack of a shared, agreed notation or ‘interchange format’ for argumentation and arguments. In this paper, we describe a draft specification for an argument interchange format (AIF) intended for representation and exchange of data between various argumentation tools and agent-based applications. It represents a consensus ‘abstract model’ established by researchers across fields of argumentation, artificial intelligence and multi-agent systems. In its current form, this specification is intended as a starting point for further discussion and elaboration by the community, rather than an attempt at a definitive, all-encompassing model. However, to demonstrate proof of concept, a use case scenario is briefly described. Moreover, three concrete realizations or ‘reifications’ of the abstract model are illustrated.

Credulous and Sceptical Argument Games for Preferred Semantics

This paper presents dialectical proof theories for Dungs preferred semantics of defeasible argumentation. The proof theories have the form of argument games for testing membership of some (credulous reasoning) or all preferred extensions (sceptical reasoning). The credulous proof theory is for the general case, while the sceptical version is for the case where preferred semantics coincides with stable semantics. The development of these argument games is especially motivated by applications of argumentation in automated negotiation, mediation of collective discussion and decision making, and intelligent tutoring.

Towards a Testbed for Multi-party Dialogues

In many situations conversations involve more than two parties. However, most research on communication modelling in e.g. multi-agent systems limits itself to conversations between two parties at a time. Very little research has been done yet on modelling multi-party dialogues. In this paper we first explore the differences between two party and multi-party dialogues and we indicate a number of issues that arise when considering dialogues between more than two parties. Then we take some steps towards creating a testbed in which these issues can be explored and theory on multi-party dialogues can be developed.

Argumentation in bayesian belief networks

This paper establishes an explicit connection between formal argumentation and Bayesian inference by introducing a notion of argument and a notion of defeat among arguments in Bayesian networks. First, the two approaches are compared and it is argued that argumentation in Bayesian belief networks is a typical multi-agent affair. Since in theories of formal argumentation the so-called admissibility semantics is an important criterion of argument validity, this paper finally proposes an algorithm to decide efficiently whether a particular node is supported by an admissible argument. The proposed algorithm is then slightly extended to an algorithm that returns the top-k of strongest admissible arguments at each node. This extension is particularly interesting from a Bayesian inference point of view, because it offers a computationally tractable alternative to the NPPP-complete decision problem k-MPE (finding the top-k most probable explanations in a Bayesian network).

Practical reasoning using values: giving meaning to values

Each person holds numerous values that represent what is believed to be important. As a result, our values influence our behavior and influence practical reasoning. Various argumentation approaches use values to justify actions, but assume knowledge about whether state transitions promote or demote values. However, this knowledge is typically disputable, since people give different meanings to the same value. This paper proposes an argumentation mechanism to argue about the meaning of an value and thus about whether state transitions promote or demote values. After giving an overview of how values are defined in social psychology, this paper defines values as preference orders and introduces several argument schemes to reason about preferences. These schemes are used to give meaning to values and to determine whether values are promoted or demoted. Furthermore, value systems are used for practical reasoning and allow resolving conflicts when pursuing your values. An example is given of how the new argument schemes can be used to do practical reasoning using values.

AVERs : an argument visualization tool for representing stories about evidence

This paper proposes an architecture for a sense-making system for crime investigation named AVERs (Argument Visualization for Evidential Reasoning based on stories). It is targeted at crime investigators who may use it to explain initially observed facts by drawing links between these facts and hypothesized events, and to connect the thus created stories to evidence through argumentation. AVERs draws on a combination of ideas from visualizing argumentation and anchored narratives theory.

Arguing about preferences and decisions

Complex decisions involve many aspects that need to be considered, which complicates determining what decision has the most preferred outcome. Artificial agents may be required to justify and discuss their decisions to others. Designers must communicate their wishes to artificial agents. Research in argumentation theory has examined how agents can argue about what decision is best using goals and values. Decisions can be justified with the goals they achieve, and goals can be justified by the values they promote. Agents may agree on having a value, but disagree about what constitutes that value. In existing work, however, it is not possible to discuss what constitutes a specific value, whether a goal promotes a value, why an agent has a value and why an agent has specific priorities over goals. This paper introduces several argument schemes, formalised in an argumentation system, to overcome these problems. The techniques presented in this paper are inspired by multi attribute decision theory. Categories and Subject Descriptors I.2.4 [Artificial Intelligence]: Knowledge Representation Formalisms and Methods. General Terms: Design.

GearBi: towards an online arbitration environment based on the design principles simplicity, awareness, orientation, and timeliness

Arbitration is a preferred method for the resolution of international business disputes. As of yet, most publications on online arbitration deal with legal issues. In this paper, we present an Online arbitration environment that we believe facilitates the participants in a meaningful way. Our assumption is that an ODR service should be easy to use (convenient), and at the same time provide meaningful support. More specifically we have paid attention to four criteria that we believe are important, viz. simplicity, awareness, orientation and timeliness. The online arbitration service is called GearBi.

A methodology for the generation of multi-agent argumentation dialogue scenarios

Increasingly research into the uses of argumentation in multi-agent dialogues takes an experimental approach. Such studies explore how agents can successfully employ argumentation besides the best and worst case situations of formal analysis. While a vital part in these experiments is influenced by the scenarios from which dialogues are generated, there is very little research on how these can be generated in a meaningful way, respecting the characteristics of the underlying dialogue problem. This paper proposes, by means of an example system for deliberation dialogues, a methodology for the construction and evaluation of a scenario generation process. It is shown how scenarios can accommodate argumentation with structured arguments and how it is tested whether the generated scenarios are interesting for experimentation.