Supervisor: Atefeh Keshavarzi(a.keshavarzi.zafarghandi@vu.nl)
One of the key advantages of symbolic AI is its explainability. However, while symbolic AI systems are theoretically explainable, providing human-understandable explanations in practice remains a significant challenge and has recently gained considerable attention. Our focus is on explaining how a particular query answer can be derived from an inconsistent dataset. A new approach has been introduced, translating inconsistent knowledge bases (KBs) into abstract dialectical frameworks (ADFs). ADFs are directed graphs where nodes represent arguments and edges reflect relationships of attack or support between arguments. In ADFs, semantics are applied to systematically determine the justification of statements at an abstract level. Certain ADF semantics have been shown to capture inconsistency-tolerant semantics for KBs. However, several open problems remain, such as how to efficiently implement this transformation, provide a solver for the newly introduced ADF semantics, and investigate their time complexity—specifically, how fast the solver performs.