Partial Preferences and Ambiguity Resolution in Contextual Defeasible Logic

Abstract

AbstractDomains, such as Ambient Intelligence and Social Networks, are characterized by some common features including distribution of the available knowledge, entities with different backgrounds, viewpoints and operational environments, and imperfect knowledge. Multi-Context Systems (MCS) has been proposed as a natural representation model for such environments, while recent studies have proposed adding non-monotonic features to MCS to address the issues of incomplete, uncertain and ambiguous information. In previous works, we introduced a non-monotonic extension to MCS and an argument-based reasoning model that handle imperfect context information based on defeasible argumentation. Here we propose alternative variants that integrate features such as partial preferences, ambiguity propagating and team defeat, and study the relations between the different variants in terms of conclusions being drawn in each case.KeywordsAmbiguity ResolutionMapping RuleAmbient IntelligenceArgumentation FrameworkPartial PreferenceThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.