Recommendation Fairness: From Static to Dynamic

Abstract

Driven by the need to capture users' evolving interests and optimize their long-term experiences, more and more recommender systems have started to model recommendation as a Markov decision process and employ reinforcement learning to address the problem. Shouldn't research on the fairness of recommender systems follow the same trend from static evaluation and one-shot intervention to dynamic monitoring and non-stop control? In this paper, we portray the recent developments in recommender systems first and then discuss how fairness could be baked into the reinforcement learning techniques for recommendation. Moreover, we argue that in order to make further progress in recommendation fairness, we may want to consider multi-agent (game-theoretic) optimization, multi-objective (Pareto) optimization, and simulationbased optimization, in the general framework of stochastic games. CCS CONCEPTS• Information systems → Recommender systems; • Computing methodologies → Reinforcement learning.

Fairness in Recommendation: A Survey

“…What's more, researches have shown that imposing static fairness criteria myopically at every step may actually exacerbate unfairness [52,54,182,214]. To solve the problem, a few works have paid attention to the dynamic factors in the recommender system environment and study how to enhance fairness with dynamics such as the change of utility, attributes and group labels due to the user interactions throughout the recommendation process [210]. Ge et al [73] study the dynamic fairness of item exposure in recommender systems.…”

“…Some researches believe that recommendation is not only a prediction problem but a sequential decision problem, and suggest to model the recommendation problem as a Markov Decision Process (MDP) and solve the problem through Reinforcement Learning (RL) [8]. Some research on fairness in recommendation also follow the trend and consider RL methods to promote fairness through a long-term and dynamic perspective [73,75,89,93,120,[180][181][182]210]. In [73], the authors consider the dynamic item exposure fairness in recommender systems where the item popularity will change over time with recommendation actions and user feedback.…”

“…One challenge of leveraging RL to promote fairness is that there may be a risk of "user tampering" [65], which is a phenomenon that a recommender system may try to increase its long-term user engagement through manipulating the user's opinions, preferences and beliefs via its recommendations. The potential unfairness of users that different users are affected by different manipulation needs vigilance from researchers and system developers [210].…”

Survey on Fair Reinforcement Learning: Theory and Practice

“…Apart from recommendation systems based on static settings, RL-based systems have provided good performance for interactive systems where users' evolving preferences must be taken into account while determining if the system is fair [144]. Liu et al [84] introduce an RL-based framework, called FairRec, that maximizes the cumulative reward function based on both fairness and accuracy, and maintains a dynamic balance between both in the long run for an interactive recommendation.…”

“…These approaches are solely based on simulations by reproducing the dynamics of certain applications with known parameters to highlight the need for modeling the long-term implications of bias. Subsequently, they have been employed in some applications such as recommendation systems to retain fairness for a longer period of time [114,144].…”