The work provides new complexity results and algorithms for reasoning over assumption-based argumentation (ABA) frameworks.

Constraint Reasoning and Optimization group will present the paper Reasoning over Assumption-Based Argumentation Frameworks via Direct Answer Set Programming Encodings, authored by Tuomo Lehtonen, Johannes P. Wallner (Vienna University of Technology, Austria), and Matti Järvisalo, at the 33rd AAAI Conference on Artificial Intelligence (AAAI-19). 

Focusing on assumption-based argumentation (ABA) as a central structured  formalism to AI argumentation, the paper presents a new approach to reasoning in ABA with and without preferences, significantly improving on the empirical performance of current state-of-the-art ABA reasoning systems. Furthermore, new complexity results for reasoning in ABA+ are provided, suggesting that the integration of preferential information into ABA results in increased problem complexity for several central argumentation semantics.

Alongside IJCAI, AAAI is the major general AI conference world-wide.