Synthetic Data Generation for Intersectional Fairness by Leveraging Hierarchical Group Structure

TL;DR

This paper presents a hierarchical data augmentation method that improves intersectional fairness in classification tasks by leveraging group structures, demonstrating superior fairness and robustness across diverse datasets.

Contribution

The paper introduces a novel hierarchical augmentation technique that enhances intersectional fairness by utilizing the inherent group structure, a significant advancement over existing methods.

Findings

Improves intersectional fairness in classifiers.

Enhances robustness against leveling down.

Effective across text and image datasets.

Abstract

In this paper, we introduce a data augmentation approach specifically tailored to enhance intersectional fairness in classification tasks. Our method capitalizes on the hierarchical structure inherent to intersectionality, by viewing groups as intersections of their parent categories. This perspective allows us to augment data for smaller groups by learning a transformation function that combines data from these parent groups. Our empirical analysis, conducted on four diverse datasets including both text and images, reveals that classifiers trained with this data augmentation approach achieve superior intersectional fairness and are more robust to ``leveling down'' when compared to methods optimizing traditional group fairness metrics.