Mitigating the disparate impact of statistical machine learning methods is
crucial for ensuring fairness. While extensive research aims to reduce
disparity, the effect of using a \emph{finite dataset} -- as opposed to the
entire population -- remains unclear. This paper explores the statistical
foundations of fair binary classification with two protected groups, focusing
on controlling demographic disparity, defined as the difference in acceptance
rates between the groups. Although fairness may come at the cost of accuracy
even with infinite data, we show that using a finite sample incurs additional
costs due to the need to estimate group-specific acceptance thresholds. We
study the minimax optimal classification error while constraining demographic
disparity to a user-specified threshold. To quantify the impact of fairness
constraints, we introduce a novel measure called \emph{fairness-aware excess
risk} and derive a minimax lower bound on this measure that all classifiers
must satisfy. Furthermore, we propose FairBayes-DDP+, a group-wise thresholding
method with an offset that we show attains the minimax lower bound. Our lower
bound proofs involve several innovations. Experiments support that
FairBayes-DDP+ controls disparity at the user-specified level, while being
faster and having a more favorable fairness-accuracy tradeoff than several
baselines.

控制统计机器学习方法的不公平影响是确保公平性的关键。本文探讨了具有两个受保护群体的公平二分类的统计基础，重点是控制人口统计差异，并提出了一种量化公平性约束影响的新方法。

有界种族差异的极小极大公平分类

Minimax Optimal Fair Classification with Bounded Demographic Disparity

The demographic disparity of biometric systems has led to serious concerns
regarding their societal impact as well as applicability of such systems in
private and public domains. A quantitative evaluation of demographic fairness
is an important step towards understanding, assessment, and mitigation of
demographic bias in biometric applications. While few, existing fairness
measures are based on post-decision data (such as verification accuracy) of
biometric systems, we discuss how pre-decision data (score distributions)
provide useful insights towards demographic fairness. In this paper, we
introduce multiple measures, based on the statistical characteristics of score
distributions, for the evaluation of demographic fairness of a generic
biometric verification system. We also propose different variants for each
fairness measure depending on how the contribution from constituent demographic
groups needs to be combined towards the final measure. In each case, the
behavior of the measure has been illustrated numerically and graphically on
synthetic data. The demographic imbalance in benchmarking datasets is often
overlooked during fairness assessment. We provide a novel weighing strategy to
reduce the effect of such imbalance through a non-linear function of sample
sizes of demographic groups. The proposed measures are independent of the
biometric modality, and thus, applicable across commonly used biometric
modalities (e.g., face, fingerprint, etc.).

本文提出了多个基于 biometric 系统评分的统计特征的评估指标和权重策略，通过评估数据得出生物识别系统的公平性。

生物识别中测量偏差的公平性指数

Fairness Index Measures to Evaluate Bias in Biometric Recognition

This work examines how to train fair classifiers in settings where training
labels are corrupted with random noise, and where the error rates of corruption
depend both on the label class and on the membership function for a protected
subgroup. Heterogeneous label noise models systematic biases towards particular
groups when generating annotations. We begin by presenting analytical results
which show that naively imposing parity constraints on demographic disparity
measures, without accounting for heterogeneous and group-dependent error rates,
can decrease both the accuracy and the fairness of the resulting classifier.
Our experiments demonstrate these issues arise in practice as well. We address
these problems by performing empirical risk minimization with carefully defined
surrogate loss functions and surrogate constraints that help avoid the pitfalls
introduced by heterogeneous label noise. We provide both theoretical and
empirical justifications for the efficacy of our methods. We view our results
as an important example of how imposing fairness on biased data sets without
proper care can do at least as much harm as it does good.

研究了在训练标签被随机噪声污染和误差率依赖于保护子组的成员函数的情况下如何训练公平分类器，发现简单地强制使用人口统计差异措施达到平等，会降低分类器的准确性和公平性。通过使用指定的替代损失函数和替代约束条件对经验风险最小化进行优化可以减少标签噪声引起的问题。