Minimax Pareto Fairness: A Multi Objective Perspective
This work addresses fairness in machine learning for applications such as healthcare and finance, offering a novel optimization perspective that is incremental in combining existing fairness concepts.
The authors tackled group fairness in classification by formulating it as a multi-objective optimization problem, where each sensitive group's risk is a separate objective, and they proposed a method achieving minimax risk and Pareto efficiency, which performed favorably in real-world case studies like income prediction and credit risk assessment.
In this work we formulate and formally characterize group fairness as a multi-objective optimization problem, where each sensitive group risk is a separate objective. We propose a fairness criterion where a classifier achieves minimax risk and is Pareto-efficient w.r.t. all groups, avoiding unnecessary harm, and can lead to the best zero-gap model if policy dictates so. We provide a simple optimization algorithm compatible with deep neural networks to satisfy these constraints. Since our method does not require test-time access to sensitive attributes, it can be applied to reduce worst-case classification errors between outcomes in unbalanced classification problems. We test the proposed methodology on real case-studies of predicting income, ICU patient mortality, skin lesions classification, and assessing credit risk, demonstrating how our framework compares favorably to other approaches.