Data Augmentation via Subgroup Mixup for Improving Fairness

TL;DR

This paper introduces a novel data augmentation technique called subgroup mixup, designed to enhance fairness in machine learning by balancing underrepresented groups and improving decision boundaries across diverse subpopulations.

Contribution

The paper presents a new pairwise mixup method for data augmentation that specifically targets fairness improvements in classification tasks, addressing societal biases and under-representation.

Findings

Achieves fair outcomes on synthetic and real-world data

Improves fairness without sacrificing accuracy

Demonstrates robustness across multiple datasets

Abstract

In this work, we propose data augmentation via pairwise mixup across subgroups to improve group fairness. Many real-world applications of machine learning systems exhibit biases across certain groups due to under-representation or training data that reflects societal biases. Inspired by the successes of mixup for improving classification performance, we develop a pairwise mixup scheme to augment training data and encourage fair and accurate decision boundaries for all subgroups. Data augmentation for group fairness allows us to add new samples of underrepresented groups to balance subpopulations. Furthermore, our method allows us to use the generalization ability of mixup to improve both fairness and accuracy. We compare our proposed mixup to existing data augmentation and bias mitigation approaches on both synthetic simulations and real-world benchmark fair classification data, demonstrating that we are able to achieve fair outcomes with robust if not improved accuracy.