This paper proposes a computational framework to reason with conflicting and gradual evidence. The framework is a synthesis of Dung’s seminal work in argumentation semantics with multi-valued logic. Abstract grounded semantics is used to identify the conditions under which a conclusion can be accepted, while multi-valued logic operators are used to quantify the degree of truth of such conditions.
We propose a truth-compositional recursive computation based on the notion of irrelevant arguments, and we discuss examples using the major multi-valued logics: Godel’s, Zadeh’s and Łukasiewicz’s logic.
The aim of this paper is to provide a sound framework for reasoning with imprecise and conflicting evidence. The core idea is to define a novel synthesis between the two main research areas relevant to the problem, namely abstract argumentation, used as a conflict-resolution strategy, and multi-valued logics, used to model gradual information.
In the argumentation setting, conclusions are reached by evaluating arguments. An argument is a construct used in discussions with a support and a claim that is derived from the support. Arguments are not proof, but rather defeasible constructs whose validity can be challenged by other arguments attacking them.
Source: Dublin Institute of Technology
Author: Pierpaolo Dondio