The multisided complexity of fairness in recommender systems

Abstract

Recommender systems are poised at the interface between stakeholders: for example, job applicants and employers in the case of recommendations of employment listings, or artists and listeners in the case of music recommendation. In such multisided platforms, recommender systems play a key role in enabling discovery of products and information at large scales. However, as they have become more and more pervasive in society, the equitable distribution of their benefits and harms have been increasingly under scrutiny, as is the case with machine learning generally. While recommender systems can exhibit many of the biases encountered in other machine learning settings, the intersection of personalization and multisidedness makes the question of fairness in recommender systems manifest itself quite differently. In this article, we discuss recent work in the area of multisided fairness in recommendation, starting with a brief introduction to core ideas in algorithmic fairness and multistakeholder recommendation. We describe techniques for measuring fairness and algorithmic approaches for enhancing fairness in recommendation outputs. We also discuss feedback and popularity effects that can lead to unfair recommendation outcomes. Finally, we introduce several promising directions for future research in this area.

Commonality in Recommender Systems: Evaluating Recommender Systems to Enhance Cultural Citizenship

“…On the other hand, consumer fairness metrics considers disparate treatments of the system in relation to different groups of consumers. Recent research has proposed more general multi-stakeholder fairness metrics, acknowledging the impact recommender systems have on different groups of individuals [18,65,87].…”

Recommender systems: Trends and frontiers

“…The works by Sonboli et al (2022) and Adomavicius et al (2022) look beyond the computer science perspective and consider the sociotechnical environment of recommender systems. Sonboli et al (2022) address the important topic of fairness in recommender systems and in particular address its multisided nature when various stakeholders should be considered. Adomavicius et al (2022), on the other hand, study the effect of "preference pollution", which may occur when the available item ratings upon which recommender system operates are biased and not representative of the true user preferences.…”

“…Moreover, they discuss the challenges that come with the evaluation of conversational systems and outline a number of future directions in the area. The works by Sonboli et al (2022) and Adomavicius et al (2022) look beyond the computer science perspective and consider the sociotechnical environment of recommender systems. Sonboli et al (2022) address the important topic of fairness in recommender systems and in particular address its multisided nature when various stakeholders should be considered.…”

“…The works by Sonboli et al. (2022) and Adomavicius et al. (2022) look beyond the computer science perspective and consider the sociotechnical environment of recommender systems.…”

“…Sonboli et al. (2022) address the important topic of fairness in recommender systems and in particular address its multisided nature when various stakeholders should be considered. Adomavicius et al.…”