Fair Ranking as Fair Division: Impact-Based Individual Fairness in Ranking

Fair Ranking as Fair Division: Impact-Based Individual Fairness in Ranking
复制标题

公平排名作为公平划分:排名中基于影响力的个人公平性

DOI:
10.1145/3534678.3539353
复制
发表时间:
2022
期刊:
ACM SIGKDD Conference on Knowledge Discovery and Data Mining
影响因子:
--
通讯作者:
Joachims, Thorsten
Joachims, Thorsten
中科院分区:
--
文献类型:
--
作者:
Saito, Yuta;Joachims, Thorsten

文献摘要

参考文献

被引文献

相似文献

排名已经成为双边在线市场的主要界面。许多人已经注意到,排名不仅影响用户的满意度(例如,客户、听众、雇主、旅行者),但是排名中的位置将曝光(从而经济机会)分配给排名项目(例如,文章、产品、歌曲、求职者、餐馆、酒店)。这就提出了公平性的问题,大多数现有的作品都明确地将项目暴露与项目相关性联系起来,以解决公平性问题。然而,我们认为,任何特定的选择这样一个链接功能可能是难以辩护,我们表明,由此产生的排名仍然是不公平的。为了避免这些缺点,我们开发了一种新的公理化方法,它植根于公平划分的原则。这不仅避免了选择链接函数的需要,而且更有意义地量化了暴露之外对项目的影响。我们的无嫉妒性公理和优于统一排名的公理假设,对于公平的排名政策,每个项目都应该优先于其他任何项目的排名分配,并且任何项目都不应该受到排名的不利影响。为了计算根据这些公理是公平的排名政策,我们提出了一个新的排名目标相关的纳什社会福利。我们表明,该解决方案有保证,它的嫉妒免费,其优势,统一排名的每一个项目,其帕累托最优。相比之下,我们发现,传统的基于安全的公平可以产生大量的嫉妒,并有一个高度不同的项目的影响。除了这些理论结果,我们实证说明我们的框架如何控制基于影响的个人项目公平性和用户效用之间的权衡。
Rankings have become the primary interface in two-sided online markets. Many have noted that the rankings not only affect the satisfaction of the users (e.g., customers, listeners, employers, travelers), but that the position in the ranking allocates exposure -- and thus economic opportunity -- to the ranked items (e.g., articles, products, songs, job seekers, restaurants, hotels). This has raised questions of fairness to the items, and most existing works have addressed fairness by explicitly linking item exposure to item relevance. However, we argue that any particular choice of such a link function may be difficult to defend, and we show that the resulting rankings can still be unfair. To avoid these shortcomings, we develop a new axiomatic approach that is rooted in principles of fair division. This not only avoids the need to choose a link function, but also more meaningfully quantifies the impact on the items beyond exposure. Our axioms of envy-freeness and dominance over uniform ranking postulate that for a fair ranking policy every item should prefer their own rank allocation over that of any other item, and that no item should be actively disadvantaged by the rankings. To compute ranking policies that are fair according to these axioms, we propose a new ranking objective related to the Nash Social Welfare. We show that the solution has guarantees regarding its envy-freeness, its dominance over uniform rankings for every item, and its Pareto optimality. In contrast, we show that conventional exposure-based fairness can produce large amounts of envy and have a highly disparate impact on the items. Beyond these theoretical results, we illustrate empirically how our framework controls the trade-off between impact-based individual item fairness and user utility.
DOI: 10.1145/3366424.3383534
发表时间: 2018-05
期刊: Companion Proceedings of the Web Conference 2020
影响因子: --
作者:
Meike Zehlike;Tom Sühr;C. Castillo;Ivan Kitanovski
通讯作者: Meike Zehlike;Tom Sühr;C. Castillo;Ivan Kitanovski
DOI: 10.1155/2014/259689
发表时间: 2014
影响因子: 4.6
作者:
Stumpfova Z;Hezova R;Meli AC;Slaby O;Michalek J
通讯作者: Michalek J
不确定性下排名的公平性
DOI: --
发表时间: 2021
期刊: Advances in neural information processing systems
影响因子: --
作者:
Singh, Ashudeep;Kempe, David;Joachims, Thorsten
通讯作者: Joachims, Thorsten
DOI: 10.4230/lipics.icalp.2018.28
发表时间: 2017-04
期刊: ArXiv
影响因子: --
作者:
L. E. Celis;D. Straszak;Nisheeth K. Vishnoi
通讯作者: L. E. Celis;D. Straszak;Nisheeth K. Vishnoi
随机强盗的暴露公平性
DOI: --
发表时间: 2021
期刊: International Conference on Machine Learning
影响因子: --
作者:
Wang, Lequn;Bai, Yiwei;Sun, Wen;Joachims, Thorsten
通讯作者: Joachims, Thorsten