Towards characterising argumentation based dialogue in the argument interchange format

  • Authors:
  • Sanjay Modgil;Jarred McGinnis

  • Affiliations:
  • Advanced Computation Lab, Cancer Research UK, London;Department of Computer Science Royal Holloway, University of London

  • Venue:
  • ArgMAS'07 Proceedings of the 4th international conference on Argumentation in multi-agent systems
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

To facilitate development and practical deployment of argumentation systems, a recent shared notation, or Argument Interchange Format (AIF), has been proposed for representation and communication of argumentation knowledge amongst agents. The AIF is described as an abstract model, or "ontology", characterising the core concepts and their relationships. Concrete reifications or syntaxes instantiating these concepts have also been described. Thus far the focus has been on representation of argument entities and networks, i.e., arguments and sub-arguments and relations of inference, preference and attack amongst these entities. Requirements were envisaged for a separate core ontology for items relating to the interchange of arguments, such as locutions and protocols. In this paper we propose that the core argument entity and network ontology can be extended to characterise communication in argumentation based dialogues between agents. We also propose a declarative specification of these communicative concepts that is of sufficient generality to serve as an operational semantics. Specifically, we propose use of the Lightweight Coordination Calculus (LCC). We then illustrate our proposal with a use case multi-agent scenario. In presenting this work, our aim is to stimulate further discussion and work on development of the AIF in order to characterise communication in multi-agent dialogues.