Harm Mitigation in Recommender Systems under User Preference Dynamics
This addresses the challenge of reducing harmful content exposure for users in dynamic recommendation environments, though it appears incremental as it builds on existing modeling and algorithmic approaches.
The paper tackles the problem of balancing click-through rate (CTR) maximization with harm mitigation in recommender systems under evolving user preferences, and it shows that their proposed policies outperform baselines in achieving this tradeoff in a semi-synthetic movie recommendation setting.
We consider a recommender system that takes into account the interplay between recommendations, the evolution of user interests, and harmful content. We model the impact of recommendations on user behavior, particularly the tendency to consume harmful content. We seek recommendation policies that establish a tradeoff between maximizing click-through rate (CTR) and mitigating harm. We establish conditions under which the user profile dynamics have a stationary point, and propose algorithms for finding an optimal recommendation policy at stationarity. We experiment on a semi-synthetic movie recommendation setting initialized with real data and observe that our policies outperform baselines at simultaneously maximizing CTR and mitigating harm.