公平测试:发现数据驱动应用程序中不合理的关联

Florian Tramèr, Vaggelis Atlidakis, Roxana Geambasu, Daniel J. Hsu, J. Hubaux, Mathias Humbert, A. Juels, Huang Lin
{"title":"公平测试:发现数据驱动应用程序中不合理的关联","authors":"Florian Tramèr, Vaggelis Atlidakis, Roxana Geambasu, Daniel J. Hsu, J. Hubaux, Mathias Humbert, A. Juels, Huang Lin","doi":"10.1109/EUROSP.2017.29","DOIUrl":null,"url":null,"abstract":"In a world where traditional notions of privacy are increasingly challenged by the myriad companies that collect and analyze our data, it is important that decision-making entities are held accountable for unfair treatments arising from irresponsible data usage. Unfortunately, a lack of appropriate methodologies and tools means that even identifying unfair or discriminatory effects can be a challenge in practice. We introduce the unwarranted associations (UA) framework, a principled methodology for the discovery of unfair, discriminatory, or offensive user treatment in data-driven applications. The UA framework unifies and rationalizes a number of prior attempts at formalizing algorithmic fairness. It uniquely combines multiple investigative primitives and fairness metrics with broad applicability, granular exploration of unfair treatment in user subgroups, and incorporation of natural notions of utility that may account for observed disparities. We instantiate the UA framework in FairTest, the first comprehensive tool that helps developers check data-driven applications for unfair user treatment. It enables scalable and statistically rigorous investigation of associations between application outcomes (such as prices or premiums) and sensitive user attributes (such as race or gender). Furthermore, FairTest provides debugging capabilities that let programmers rule out potential confounders for observed unfair effects. We report on use of FairTest to investigate and in some cases address disparate impact, offensive labeling, and uneven rates of algorithmic error in four data-driven applications. As examples, our results reveal subtle biases against older populations in the distribution of error in a predictive health application and offensive racial labeling in an image tagger.","PeriodicalId":233564,"journal":{"name":"2017 IEEE European Symposium on Security and Privacy (EuroS&P)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2015-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"146","resultStr":"{\"title\":\"FairTest: Discovering Unwarranted Associations in Data-Driven Applications\",\"authors\":\"Florian Tramèr, Vaggelis Atlidakis, Roxana Geambasu, Daniel J. Hsu, J. Hubaux, Mathias Humbert, A. Juels, Huang Lin\",\"doi\":\"10.1109/EUROSP.2017.29\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In a world where traditional notions of privacy are increasingly challenged by the myriad companies that collect and analyze our data, it is important that decision-making entities are held accountable for unfair treatments arising from irresponsible data usage. Unfortunately, a lack of appropriate methodologies and tools means that even identifying unfair or discriminatory effects can be a challenge in practice. We introduce the unwarranted associations (UA) framework, a principled methodology for the discovery of unfair, discriminatory, or offensive user treatment in data-driven applications. The UA framework unifies and rationalizes a number of prior attempts at formalizing algorithmic fairness. It uniquely combines multiple investigative primitives and fairness metrics with broad applicability, granular exploration of unfair treatment in user subgroups, and incorporation of natural notions of utility that may account for observed disparities. We instantiate the UA framework in FairTest, the first comprehensive tool that helps developers check data-driven applications for unfair user treatment. It enables scalable and statistically rigorous investigation of associations between application outcomes (such as prices or premiums) and sensitive user attributes (such as race or gender). Furthermore, FairTest provides debugging capabilities that let programmers rule out potential confounders for observed unfair effects. We report on use of FairTest to investigate and in some cases address disparate impact, offensive labeling, and uneven rates of algorithmic error in four data-driven applications. As examples, our results reveal subtle biases against older populations in the distribution of error in a predictive health application and offensive racial labeling in an image tagger.\",\"PeriodicalId\":233564,\"journal\":{\"name\":\"2017 IEEE European Symposium on Security and Privacy (EuroS&P)\",\"volume\":\"22 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2015-10-08\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"146\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2017 IEEE European Symposium on Security and Privacy (EuroS&P)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/EUROSP.2017.29\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2017 IEEE European Symposium on Security and Privacy (EuroS&P)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/EUROSP.2017.29","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 146

摘要

在一个传统的隐私观念日益受到无数收集和分析我们数据的公司挑战的世界里,重要的是,决策实体要对不负责任的数据使用所带来的不公平待遇负责。不幸的是,缺乏适当的方法和工具意味着,即使确定不公平或歧视性影响,在实践中也可能是一项挑战。我们介绍了无担保关联(UA)框架,这是一种在数据驱动的应用程序中发现不公平、歧视性或冒犯性用户待遇的原则性方法。UA框架统一并合理化了之前的一些形式化算法公平性的尝试。它独特地结合了多种调查原语和公平指标,具有广泛的适用性,对用户子组中的不公平待遇进行了细致的探索,并结合了可能解释观察到的差异的效用的自然概念。我们在FairTest中实例化了UA框架,这是第一个全面的工具,可以帮助开发人员检查数据驱动的应用程序是否存在不公平的用户待遇。它支持对应用程序结果(如价格或保费)和敏感用户属性(如种族或性别)之间的关联进行可扩展和统计上严格的调查。此外,FairTest提供了调试功能,使程序员能够排除观察到的不公平影响的潜在混杂因素。我们报告了在四个数据驱动的应用程序中使用FairTest来调查并在某些情况下解决不同的影响、攻击性标签和不均匀的算法错误率。例如,我们的研究结果揭示了在预测健康应用程序的错误分布中对老年人的微妙偏见,以及在图像标记器中对冒犯性种族标签的偏见。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
FairTest: Discovering Unwarranted Associations in Data-Driven Applications
In a world where traditional notions of privacy are increasingly challenged by the myriad companies that collect and analyze our data, it is important that decision-making entities are held accountable for unfair treatments arising from irresponsible data usage. Unfortunately, a lack of appropriate methodologies and tools means that even identifying unfair or discriminatory effects can be a challenge in practice. We introduce the unwarranted associations (UA) framework, a principled methodology for the discovery of unfair, discriminatory, or offensive user treatment in data-driven applications. The UA framework unifies and rationalizes a number of prior attempts at formalizing algorithmic fairness. It uniquely combines multiple investigative primitives and fairness metrics with broad applicability, granular exploration of unfair treatment in user subgroups, and incorporation of natural notions of utility that may account for observed disparities. We instantiate the UA framework in FairTest, the first comprehensive tool that helps developers check data-driven applications for unfair user treatment. It enables scalable and statistically rigorous investigation of associations between application outcomes (such as prices or premiums) and sensitive user attributes (such as race or gender). Furthermore, FairTest provides debugging capabilities that let programmers rule out potential confounders for observed unfair effects. We report on use of FairTest to investigate and in some cases address disparate impact, offensive labeling, and uneven rates of algorithmic error in four data-driven applications. As examples, our results reveal subtle biases against older populations in the distribution of error in a predictive health application and offensive racial labeling in an image tagger.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信