On The Fairness Impacts of Hardware Selection in Machine Learning
This addresses fairness issues in ML-as-a-service platforms where users lack hardware control, highlighting an overlooked aspect with potential broad implications.
The paper tackles the problem of how hardware selection in machine learning affects model fairness, showing that hardware choices can worsen disparities across demographic groups due to variations in gradient flows and loss surfaces, and proposes a mitigation strategy.
In the machine learning ecosystem, hardware selection is often regarded as a mere utility, overshadowed by the spotlight on algorithms and data. This oversight is particularly problematic in contexts like ML-as-a-service platforms, where users often lack control over the hardware used for model deployment. How does the choice of hardware impact generalization properties? This paper investigates the influence of hardware on the delicate balance between model performance and fairness. We demonstrate that hardware choices can exacerbate existing disparities, attributing these discrepancies to variations in gradient flows and loss surfaces across different demographic groups. Through both theoretical and empirical analysis, the paper not only identifies the underlying factors but also proposes an effective strategy for mitigating hardware-induced performance imbalances.