Achieving User-Side Fairness in Contextual Bandits

Achieving User-Side Fairness in Contextual Bandits
复制标题

在上下文强盗中实现用户端公平

DOI:
10.1007/s44230-022-00008-w
复制
发表时间:
2022
期刊:
Human-Centric Intelligent Systems
影响因子:
--
通讯作者:
Heffernan, Neil
Heffernan, Neil
中科院分区:
--
文献类型:
--
作者:
Huang, Wen;Labille, Kevin;Wu, Xintao;Lee, Dongwon;Heffernan, Neil

文献摘要

参考文献

被引文献

相似文献

基于多臂强盗(MAB)算法的个性化推荐可以根据反馈动态调整推荐策略,具有较高的实用性和效率。然而,在个性化推荐中可能会产生不公平。本文主要研究如何在个性化推荐中实现用户端的公平性。我们将我们的公平个性化推荐作为一个修改后的上下文强盗,并专注于在被推荐的物品上实现公平,而不是在被推荐的物品上实现公平。我们引入并定义了一个指标,该指标根据特权群体和受保护群体获得的奖励来衡量公平性。本文在传统的LinUCB算法的基础上,提出了一种公平上下文强盗算法fair -LinUCB,实现了用户的组级公平。我们的算法检测和监控不公平,同时学习向学生推荐个性化视频,以实现高效率。我们提供了一个理论上的遗憾分析,并表明我们的算法比LinUCB具有略高的遗憾界。我们进行了大量的实验评估,以比较我们的公平上下文强盗与LinUCB的性能,并表明我们的方法在保持高效用的同时实现了群体层面的公平。
Personalized recommendation based on multi-arm bandit (MAB) algorithms has shown to lead to high utility and efficiency as it can dynamically adapt the recommendation strategy based on feedback. However, unfairness could incur in personalized recommendation. In this paper, we study how to achieve user-side fairness in personalized recommendation. We formulate our fair personalized recommendation as a modified contextual bandit and focus on achieving fairness on the individual whom is being recommended an item as opposed to achieving fairness on the items that are being recommended. We introduce and define a metric that captures the fairness in terms of rewards received for both the privileged and protected groups. We develop a fair contextual bandit algorithm, Fair-LinUCB, that improves upon the traditional LinUCB algorithm to achieve group-level fairness of users. Our algorithm detects and monitors unfairness while it learns to recommend personalized videos to students to achieve high efficiency. We provide a theoretical regret analysis and show that our algorithm has a slightly higher regret bound than LinUCB. We conduct numerous experimental evaluations to compare the performances of our fair contextual bandit to that of LinUCB and show that our approach achieves group-level fairness while maintaining a high utility.
DOI: 10.1145/3269206.3271795
发表时间: 2018-10
期刊: Proceedings of the 27th ACM International Conference on Information and Knowledge Management
影响因子: --
作者:
Ziwei Zhu;Xia Hu;James Caverlee
通讯作者: Ziwei Zhu;Xia Hu;James Caverlee
使用基于 Bandit 的受限个性化平衡新闻
DOI: --
发表时间: 2018
期刊: International Joint Conference on Artificial Intelligence
影响因子: --
作者:
Sayash Kapoor;Sayash Kapoor;Vijay Keswani;Nisheeth K. Vishnoi;L. E. Celis
通讯作者: L. E. Celis
DOI: 10.1145/nnnnnnn.nnnnnnn
发表时间: 2017-05
期刊: 2022 IEEE/ACM 30th International Conference on Program Comprehension (ICPC)
影响因子: --
作者:
Yang Liu;Goran Radanovic;Christos Dimitrakakis;Debmalya Mandal;D. Parkes
通讯作者: Yang Liu;Goran Radanovic;Christos Dimitrakakis;Debmalya Mandal;D. Parkes
DOI: 10.1353/jhr.2009.0024
发表时间: 2009-03-01
影响因子: 5.2
作者:
Hoffmann, Florian;Oreopoulos, Philip
通讯作者: Oreopoulos, Philip
具有高概率公平保证的离线上下文强盗
DOI: --
发表时间: 2019
期刊: Advances in neural information processing systems
影响因子: --
作者:
Metevier, Blossom;Giguere, Stephen;Brockman, Sarah;Kobren, Ari;Brun, Yuriy;Brunskill, Emma;Thomas, Philip
通讯作者: Thomas, Philip