Operationalizing Individual Fairness with Pairwise Fair Representations
Operationalizing Individual Fairness with Pairwise Fair Representations
复制标题
通过配对公平表示实现个人公平
DOI：
--
复制
发表时间：
2019
影响因子：
2.5
通讯作者：
G. Weikum
中科院分区：
文献类型：
--
作者：
Preethi Lahoti;K. Gummadi;G. Weikum
我们重新审视Dwork等人提出的个人公平的概念。在操作他们的方法的一个核心挑战是难以引出一个相似度量的人类规范。在本文中，我们提出了一种不依赖于距离度量的人类规范的个人公平的操作化。相反，我们提出了新的方法来引出和利用同样值得的个人的侧面信息，以对抗社会群体之间的从属关系。我们将这些知识建模为公平图，并学习数据的统一成对公平表示（PFR），该数据既捕获数据驱动的个体之间的相似性，也捕获公平图中的成对侧信息。我们从各种来源引出公平性判断，包括人类对两个现实世界数据集的再犯预测（COMPAS）和暴力社区预测（Crime & Communities）的判断。我们的实验表明，PFR模型在实践中是可行的。
We revisit the notion of individual fairness proposed by Dwork et al. A central challenge in operationalizing their approach is the difficulty in eliciting a human specification of a similarity metric. In this paper, we propose an operationalization of individual fairness that does not rely on a human specification of a distance metric. Instead, we propose novel approaches to elicit and leverage side-information on equally deserving individuals to counter subordination between social groups. We model this knowledge as a fairness graph, and learn a unified Pairwise Fair Representation (PFR) of the data that captures both data-driven similarity between individuals and the pairwise side-information in fairness graph. We elicit fairness judgments from a variety of sources, including human judgments for two real-world datasets on recidivism prediction (COMPAS) and violent neighborhood prediction (Crime & Communities). Our experiments show that the PFR model for operationalizing individual fairness is practically viable.
DOI：
10.5441/002/edbt.2018.22
发表时间：
2018
期刊：
--
影响因子：
--
作者：
Julia Stoyanovich;Ke Yang;H. V. Jagadish
通讯作者：
Julia Stoyanovich;Ke Yang;H. V. Jagadish