Balancing Beyond Discrete Categories: Continuous Demographic Labels for Fair Face Recognition

TL;DR

This paper proposes using continuous demographic labels instead of discrete categories to better understand and mitigate bias in face recognition, demonstrating improved model fairness and performance.

Contribution

It introduces a novel continuous demographic labeling approach and validates its effectiveness both experimentally and theoretically.

Findings

Models trained on continuous-balanced data outperform discrete-balanced models.

Not all identities within the same ethnicity contribute equally to dataset balance.

Continuous demographic labels provide deeper insight into bias mitigation.

Abstract

Bias has been a constant in face recognition models. Over the years, researchers have looked at it from both the model and the data point of view. However, their approach to mitigation of data bias was limited and lacked insight on the real nature of the problem. Here, in this document, we propose to revise our use of ethnicity labels as a continuous variable instead of a discrete value per identity. We validate our formulation both experimentally and theoretically, showcasing that not all identities from one ethnicity contribute equally to the balance of the dataset; thus, having the same number of identities per ethnicity does not represent a balanced dataset. We further show that models trained on datasets balanced in the continuous space consistently outperform models trained on data balanced in the discrete space. We trained more than 65 different models, and created more than 20 subsets of the original datasets.