The Connection Between Popularity Bias, Calibration, and Fairness in Recommendation

被引:67
作者
Abdollahpouri, Himan [1 ]
Mansoury, Masoud [2 ]
Burke, Robin [1 ]
Mobasher, Bamshad [3 ]
机构
[1] Univ Colorado, Boulder, CO 80309 USA
[2] Eindhoven Univ Technol, Eindhoven, Netherlands
[3] Depaul Univ, Chicago, IL 60604 USA
来源
RECSYS 2020: 14TH ACM CONFERENCE ON RECOMMENDER SYSTEMS | 2020年
基金
美国国家科学基金会;
关键词
Recommender systems; Calibration; Algorithmic bias; Popularity bias amplification;
D O I
10.1145/3383313.3418487
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Recently there has been a growing interest in fairness-aware recommender systems including fairness in providing consistent performance across different users or groups of users. A recommender system could be considered unfair if the recommendations do not fairly represent the tastes of a certain group of users while other groups receive recommendations that are consistent with their preferences. In this paper, we use a metric called miscalibration for measuring how a recommendation algorithm is responsive to users' true preferences and we consider how various algorithms may result in different degrees of miscalibration for different users. In particular, we conjecture that popularity bias which is a wellknown phenomenon in recommendation is one important factor leading to miscalibration in recommendation. Our experimental results using two real-world datasets show that there is a connection between how different user groups are affected by algorithmic popularity bias and their level of interest in popular items. Moreover, we show that the more a group is affected by the algorithmic popularity bias, the more their recommendations are miscalibrated.
引用
收藏
页码:726 / 731
页数:6
相关论文
共 31 条
[1]   Multistakeholder recommendation: Survey and research directions [J].
Abdollahpouri, Himan ;
Adomavicius, Gediminas ;
Burke, Robin ;
Guy, Ido ;
Jannach, Dietmar ;
Kamishima, Toshihiro ;
Krasnodebski, Jan ;
Pizzato, Luiz .
USER MODELING AND USER-ADAPTED INTERACTION, 2020, 30 (01) :127-158
[2]  
Abdollahpouri Himan, 2019, 32 INT FLOR ART INT
[3]   Improving Aggregate Recommendation Diversity Using Ranking-Based Techniques [J].
Adomavicius, Gediminas ;
Kwon, YoungOk .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2012, 24 (05) :896-911
[4]  
Anderson S., 2006, LONG TAIL WHY FUTURE
[5]  
[Anonymous], 2017, ADV NEUR IN
[6]   Bias in algorithmic filtering and personalization [J].
Bozdag, Engin .
ETHICS AND INFORMATION TECHNOLOGY, 2013, 15 (03) :209-227
[7]  
Brynjolfsson E, 2006, MIT SLOAN MANAGE REV, V47, P67
[8]  
Dwork C., 2012, P 3 INN THEOR COMP
[9]  
Ekstrand Michael D, 2018, P 1 C FAIRN ACC TRAN, P172
[10]  
Garvie C., 2016, The Atlantic