Causal Argumentation Framework for Explainable AI

A new method integrates causality with argument-based reasoning to explain machine learning predictions. The approach uses causal discovery to identify relationships among variables, then translates them into a Bipolar Argumentation Framework (BAF) representing supportive and opposing feature interactions. Semi-stable semantics find extensions that explain outcomes. The method is demonstrated on two benchmark datasets and compared against standard post-hoc explainability approaches.