Feb 1, 2023 · We propose a fair classification algorithm which can satisfy the group fairness constraints with finite-sample and distribution-free guarantees.

In this paper, we propose a post-processing algorithm FaiREE that provably achieves group fairness guarantees with only finite-sample and free of distributional assumptions (this property …

To address these concerns, drawing inspiration from FaiREE (Li et al., 2022), a post-processing method designed for achieving fairness in finite-sample and distribution-free scenarios, this …

Abstract Deep neural networks are prone to various bias issues, jeopardizing their applications for high-stake decision-making. Existing fairness methods typically ofer a fixed accuracy-fairness …

YouOnlyDebiasOnce:TowardsFlexible Accuracy-FairnessTrade-offsatInferenceTime You Only Debias Once: Towards Flexible Accuracy-Fairness Trade-ofs at Inference Time

Abstract Fair representation learning (FRL) is a popular class of methods aiming to produce fair classi-fiers via data preprocessing. Recent regulatory directives stress the need for FRL …