Modelling argumentation lines in abstract frameworks

• Autores: Martínez, Diego C.; Simari, Guillermo Ricardo
• Año de publicación: 2005
• Idioma: inglés
• Tipo de recurso: documento de conferencia
• Materia: Ciencias Informáticas; modelling argumentation lines; abstract frameworks; ARTIFICIAL INTELLIGENCE; Frameworks
• Institución: Universidad Nacional de La Plata

The area of Knowledge Representation and Reasoning has been enriched during the past two decades with the addition of Argument-Based Reasoning Systems. In formal systems of defeasible argumentation, arguments for and against a proposition are produced and evaluated to test the acceptability of that proposition following a dialectical process. The main idea in these systems is that a proposition will be accepted as true if there exists an argument that supports it, and this argument is acceptable according to an analysis between it and its counterarguments (i.e., conflictive arguments). This analysis requires a process of comparison of conflicting arguments in order to decide which one is preferable. After this dialectical analysis is performed over the set of arguments in the system, some of them will be acceptable or justified arguments, while others will not. Argumentation is used as a form of non-monotonic or defeasible reasoning and it is suitable for modeling dialogues between intelligent agents. Abstract argumentation is a pathway to the study of common properties of defeasible argumentation systems. Due to the wide range of proposals in this area, each one introducing a new logic and therefore a new argument architecture, it is very interesting to define what may be understood as a correct process of argumentation, despites the differences among these systems. Perhaps one of the most important frameworks for abstract argumentation is defined by Dung in. In this work, arguments are abstract entities and no reference to the underlying logic is made. Arguments can attack other arguments, so a binary relation in the set of arguments is included, in order to model the set of conflicts in the framework. Several argument extensions capture the semantic of this system. The beauty of this work resides on its simplicity. However, some elements are missing, such as a subargument relation, and therefore some controversial situations can not be modeled.
Eje: Inteligencia artificial
Red de Universidades con Carreras en Informática (RedUNCI)