PhD project: Evaluating the impact of defeasible argumentation theory for inference under uncertainty
Argumentation theory (AT) is an important area of logic-based artificial intelligence, which provides the basis for computational models of defeasible reasoning. Despite promising progresses have been made in several areas, demonstrating AT as a solid theoretical research discipline for implementing defeasible reasoning in practice, there are issues for applied research. State-of-the-art models of AT are usually domain dependent, not often built upon all the layers of an argumentative process. Due to this diversity, a clear structure that can be replicated and that allows models to be designed, built, evaluated and compared has not emerged yet. The aim of this research is to design an argument-based framework (ABF), from the construction of arguments, to the resolution of possible inconsistencies arising from their interactions and the computation of a justifiable conclusion or claim. This ABF is proposed to be evaluated across practical applications in the fields of knowledge representation and decision-making. In this study it is believed that since AT is a relatively new field the proposal of a more generally applicable solution, in the form of a computational framework, might facilitate comparisons across applications end enable the demonstration of the impact of defeasible reasoning.
Timeframe
Funding
Lead PI(s)
Dr. Luca Longo