Bringing Friends into the Loop of Recommender Systems: an Exploratory Study.
Jinghua Piao,Guozhen Zhang,Fengli Xu,Zhilong Chen,Yu Zheng,Chen Gao,Yong Li
DOI: https://doi.org/10.1145/3479583
2021-01-01
Proceedings of the ACM on Human-Computer Interaction
Abstract:The recommender system (RS), as a computer-supported information filtering system, is ubiquitous and influences what we eat, watch, or even like. In online RS, interactions between users and the system form a feedback loop: users take actions based on the recommendations provided by RS, and RS updates its recommendations accordingly. As such interactions increase, the issue of recommendation homogeneity intensifies, which significantly impairs user experience. In the face of this long-standing issue, the newly-emerging social e-commerce offers a new solution -- bringing friends' recommendations into the loop (friend-in-the-loop). In this paper, we conduct an exploratory study on the benefits of friend-in-the-loop through mixed methods on a leading social e-commerce platform in China, Beidian. We reveal that friend-in-the-loop provides users with more accurate and diverse recommendations than merely RS, and significantly alleviates algorithmic homogeneity. Moreover, our qualitative results demonstrate that the introduction of friends' external knowledge, consumers' trust, and empathy accounts for these benefits. Overall, we elaborate that friend-in-the-loop comprehensively benefits both users and RS, and it is a promising HCI-based solution to recommendation homogeneity, which offers insightful implications on designing future human-algorithm collaboration models.