Arbitrariness Lies Beyond the Fairness-Accuracy Frontier
This work highlights a critical flaw in current fairness practices for machine learning applications affecting individuals, such as decision-making systems, by showing that interventions may increase arbitrariness, thus impacting fairness and reliability in high-stakes domains.
The paper demonstrates that fairness interventions in machine learning, focused on group fairness and accuracy, can worsen predictive multiplicity, where models achieve similar performance but produce conflicting outputs for individual samples, potentially masking high arbitrariness behind favorable metrics. It proposes an ensemble algorithm that provably ensures more consistent predictions to address this issue.
Machine learning tasks may admit multiple competing models that achieve similar performance yet produce conflicting outputs for individual samples -- a phenomenon known as predictive multiplicity. We demonstrate that fairness interventions in machine learning optimized solely for group fairness and accuracy can exacerbate predictive multiplicity. Consequently, state-of-the-art fairness interventions can mask high predictive multiplicity behind favorable group fairness and accuracy metrics. We argue that a third axis of ``arbitrariness'' should be considered when deploying models to aid decision-making in applications of individual-level impact. To address this challenge, we propose an ensemble algorithm applicable to any fairness intervention that provably ensures more consistent predictions.