Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.782068
Title: Meta-argumentation frameworks for modelling dialogues with information from sources of varying trustworthiness
Author: Ogunniye, Gideon
ISNI:       0000 0004 7967 674X
Awarding Body: University of Aberdeen
Current Institution: University of Aberdeen
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
In multi-party dialogues, arguments may originate from several sources with varying degrees of trustworthiness, for example due to maliciousness, incompetence or the like. In such dialogues, uncertainties about the trustworthiness of arguments and their sources will impact the strength attributed to the arguments, and will in turn affect the reasoning process. In human dialogues, such sources are characterised by the lack of trust ascribed to them. To obtain appropriate conclusions, trust reasoning should be integrated into formal frameworks for dialogues to identify and extract the most trustworthy conclusions. A dialogue framework should be robust to consider trust in arguments and their sources so as to minimise inconsistencies in the information obtained from a dialogue. Recently, a number of argumentation-based frameworks for dialogues have been developed to consider trust when defining preferences between arguments. However, existing frameworks have not considered how and why trust in arguments may change within a dialogue and how such change in trust should affect the justified conclusions of the dialogue. In this research, we propose meta-argumentation frameworks for multi-party dialogues in which participants consider how much they trust each other and the advanced arguments in order to define their preferences over the arguments, given that arguments (or information that supports the arguments) from more trustworthy sources may be preferred to arguments from less trustworthy sources. Dialogue participants may have conflicting preferences over arguments based on their subjective views about trust. We describe preference aggregation technique to resolve such conflicting preferences. Within our frameworks, we propose a number of argument schemes for meta-level reasoning about trust in arguments and their sources that can be integrated into any dialectical argumentation framework. These argument schemes encode intuitive properties for the dynamic trust ratings of arguments and their sources. We provide the formal analysis of our frameworks to establish how dynamic trust ratings of arguments should affect the justified conclusions of a dialogue. We also establish how dynamic trust ratings of arguments and their sources can be integrated into the strategies of dialogue participants to make suitable moves towards achieving their dialogical goals. We explore the use of dynamic trust ratings of arguments leading to preference orderings over them as a strategy for conflict resolution in argumentation-based dialogues. Through the proposed meta-argumentation frameworks, we advance the state-of-the-art in dialectical argumentation.
Supervisor: Oren, Nir Sponsor: Tertiary Education Trust Fund (Nigeria)
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.782068  DOI: Not available
Keywords: Dialogue ; Argumentation
Share: