fax-ai 0.1.0

A explanation based approach for fair supervised learning

Fair and Explainable AI (FaX-AI) framework provides implementations for fair learning methods that remove direct discrimination without the induction of indirect discrimination. These methods are based on a joining of concepts in fairness and explainability literature in machine learning. They inhibit discrimination by nullifying the influence of the protected feature on a system’s output, while preserving the influence of remaining features.