Fair Mixup: Fairness via Interpolation
This addresses fairness generalization issues in machine learning, offering a method to enhance fairness and accuracy in various domains, though it appears incremental as it builds on existing mixup techniques.
The paper tackled the problem of fair classifiers not generalizing well at evaluation time despite satisfying fairness constraints during training, and proposed fair mixup, a data augmentation strategy that improves generalization for both accuracy and fairness across tabular, vision, and language benchmarks.
Training classifiers under fairness constraints such as group fairness, regularizes the disparities of predictions between the groups. Nevertheless, even though the constraints are satisfied during training, they might not generalize at evaluation time. To improve the generalizability of fair classifiers, we propose fair mixup, a new data augmentation strategy for imposing the fairness constraint. In particular, we show that fairness can be achieved by regularizing the models on paths of interpolated samples between the groups. We use mixup, a powerful data augmentation strategy to generate these interpolates. We analyze fair mixup and empirically show that it ensures a better generalization for both accuracy and fairness measurement in tabular, vision, and language benchmarks.