Fairness of Deep Ensembles: On the interplay between per-group task difficulty and under-representation

TL;DR

This paper investigates how deep ensemble methods can improve fairness across demographic groups by mitigating disparities without sacrificing overall accuracy, and examines the complex effects of subgroup representation and task difficulty.

Contribution

It demonstrates that simple homogeneous ensembles can enhance fairness for underperforming groups and explores the nuanced impact of subgroup balance and task difficulty on model bias.

Findings

Ensembles can reduce disparities without harming overall performance.

Balanced datasets may be suboptimal when subgroup task difficulty varies.

Task difficulty and subgroup representation interact complexly affecting fairness.

Abstract

Ensembling is commonly regarded as an effective way to improve the general performance of models in machine learning, while also increasing the robustness of predictions. When it comes to algorithmic fairness, heterogeneous ensembles, composed of multiple model types, have been employed to mitigate biases in terms of demographic attributes such as sex, age or ethnicity. Moreover, recent work has shown how in multi-class problems even simple homogeneous ensembles may favor performance of the worst-performing target classes. While homogeneous ensembles are simpler to implement in practice, it is not yet clear whether their benefits translate to groups defined not in terms of their target class, but in terms of demographic or protected attributes, hence improving fairness. In this work we show how this simple and straightforward method is indeed able to mitigate disparities, particularly benefiting under-performing subgroups. Interestingly, this can be achieved without sacrificing overall performance, which is a common trade-off observed in bias mitigation strategies. Moreover, we analyzed the interplay between two factors which may result in biases: sub-group under-representation and the inherent difficulty of the task for each group. These results revealed that, contrary to popular assumptions, having balanced datasets may be suboptimal if the task difficulty varies between subgroups. Indeed, we found that a perfectly balanced dataset may hurt both the overall performance and the gap between groups. This highlights the importance of considering the interaction between multiple forces at play in fairness.