Minimax Pareto Fairness: A Multi Objective Perspective

Minimax Pareto Fairness: A Multi Objective Perspective
复制标题

DOI:
--
复制
发表时间:
2020-07
期刊:
Proceedings of machine learning research
影响因子:
--
通讯作者:
Natalia Martínez;Martín Bertrán;G. Sapiro
Natalia Martínez;Martín Bertrán;G. Sapiro
中科院分区:
其他
文献类型:
--
作者:
Natalia Martínez;Martín Bertrán;G. Sapiro

文献摘要

被引文献

相似文献

在这项工作中,我们将群体的公平性表征为多目标优化问题,其中每个敏感的群体风险都是一个单独的目标。我们提出了一个公平标准,分类器会达到最小风险,并且是帕累托高效的W.R.T.所有群体都避免了不必要的伤害,如果政策决定了这一点,则可以导致最佳的零差距模型。我们提供了一种简单的优化算法与深神网络兼容的算法,以满足这些约束。由于我们的方法不需要测试时间访问敏感属性,因此可以将其应用于减少不平衡分类问题结果之间的最坏情况分类错误。我们在预测收入,ICU患者死亡率,皮肤病变分类和评估信用风险的实际病例研究中测试了提出的方法,并证明了我们的框架与其他方法的有利相比。
In this work we formulate and formally characterize group fairness as a multi-objective optimization problem, where each sensitive group risk is a separate objective. We propose a fairness criterion where a classifier achieves minimax risk and is Pareto-efficient w.r.t. all groups, avoiding unnecessary harm, and can lead to the best zero-gap model if policy dictates so. We provide a simple optimization algorithm compatible with deep neural networks to satisfy these constraints. Since our method does not require test-time access to sensitive attributes, it can be applied to reduce worst-case classification errors between outcomes in unbalanced classification problems. We test the proposed methodology on real case-studies of predicting income, ICU patient mortality, skin lesions classification, and assessing credit risk, demonstrating how our framework compares favorably to other approaches.