Counterfactual Interventions

This paper presents a novel framework, Concept Influence for Fairness (CIF), to understand and address the issue of unfairness in machine learning models. The authors propose a method to quantify…