Differentially Private Fair Learning
Differentially Private Fair Learning
复制标题
DOI：
--
复制
发表时间：
2018-12
期刊：
影响因子：
--
通讯作者：
Matthew Jagielski;Michael Kearns;Jieming Mao;Alina Oprea;Aaron Roth;Saeed Sharifi-Malvajerdi;Jonathan Ullman
中科院分区：
文献类型：
--
作者：
Matthew Jagielski;Michael Kearns;Jieming Mao;Alina Oprea;Aaron Roth;Saeed Sharifi-Malvajerdi;Jonathan Ullman
由于预测模型可能被要求在某些属性（如种族）方面不具有歧视性，但即使收集敏感属性也可能被禁止或限制，因此我们启动了差分隐私约束下的公平学习研究。我们设计了两种学习算法，同时保证了差异隐私和均等几率，这是一种“公平”条件，对应于在受保护的群体中均衡误报率和负率。我们的第一个算法是[Hardt等人，2016]的等赔率后处理方法的私有实现。这个算法非常简单，但必须能够在测试时明确地使用受保护的群体成员，这可以被视为一种“区别对待”的形式。我们的第二种算法是[Agarwal等人，2018]的oracle-efficient in-processing方法的差异私有版本，可以用来找到最优的公平分类器，给定一个可以解决原始（不一定公平）学习问题的子程序。该算法较为复杂，但在测试时不需要访问保护组成员。我们确定了公平性、准确性和隐私之间的新权衡，这些权衡只有在需要所有这三个属性时才会出现，并表明如果在测试时使用组成员关系，这些权衡可以更温和。最后我们做了一个简短的实验评价。
Motivated by settings in which predictive models may be required to be non-discriminatory with respect to certain attributes (such as race), but even collecting the sensitive attribute may be forbidden or restricted, we initiate the study of fair learning under the constraint of differential privacy. We design two learning algorithms that simultaneously promise differential privacy and equalized odds, a 'fairness' condition that corresponds to equalizing false positive and negative rates across protected groups. Our first algorithm is a private implementation of the equalized odds post-processing approach of [Hardt et al., 2016]. This algorithm is appealingly simple, but must be able to use protected group membership explicitly at test time, which can be viewed as a form of 'disparate treatment'. Our second algorithm is a differentially private version of the oracle-efficient in-processing approach of [Agarwal et al., 2018] that can be used to find the optimal fair classifier, given access to a subroutine that can solve the original (not necessarily fair) learning problem. This algorithm is more complex but need not have access to protected group membership at test time. We identify new tradeoffs between fairness, accuracy, and privacy that emerge only when requiring all three properties, and show that these tradeoffs can be milder if group membership may be used at test time. We conclude with a brief experimental evaluation.