mlsanityData to 7 Oct 2026

Harm Mitigation in Recommender Systems under User Preference Dynamics

arXiv 2406.09882 ↗First posted 14 Jun 2024cs.IR cs.CY cs.LGcited by 5

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 25 of its 79 references that are in the map. First posted on arXiv 14 Jun 2024.

Builds on

25 of 79 references.

References · 25

  1. 2021Efficient and Modular Implicit Differentiationcited by 199
  2. 2020Feedback Loop and Bias Amplification in Recommender Systemscited by 89
  3. 2019Degenerate Feedback Loops in Recommender Systemscited by 67
  4. 2019Auditing Radicalization Pathways on YouTubecited by 52
  5. 2020Understanding Echo Chambers in E-commerce Recommender Systemscited by 40
  6. 2021Algorithmic Amplification of Politics on Twittercited by 38
  7. 2017Thompson Sampling for the MNL-Banditcited by 35
  8. 2017Gradient Methods for Submodular Maximizationcited by 33
  9. 2021Detecting Harmful Content On Online Platforms: What Platforms Need Vs. Where Research Efforts Gocited by 32
  10. 2022Counterfactual harmcited by 29
  11. 2020Do Offline Metrics Predict Online Performance in Recommender Systems?cited by 25
  12. 2022Preference Dynamics Under Personalized Recommendationscited by 25
13 more
  1. 2023An Empirical Study of Metrics to Measure Representational Harms in Pre-Trained Language Modelscited by 24
  2. 2021Multinomial Logit Contextual Bandits: Provable Optimality and Practicalitycited by 21
  3. 2022A Causal Analysis of Harmcited by 15
  4. 2018The closed loop between opinion formation and personalised recommendationscited by 12
  5. 2023A Classification of Feedback Loops and Their Relation to Biases in Automated Decision-Making Systemscited by 12
  6. 2018Modelling Opinion Dynamics in the Age of Algorithmic Personalisationcited by 9
  7. 2022Towards Psychologically-Grounded Dynamic Preference Modelscited by 8
  8. 2022How does Twitter account moderation work? Dynamics of account creation and suspension on Twitter during major geopolitical eventscited by 6
  9. 2022Rewiring What-to-Watch-Next Recommendations to Reduce Radicalization Pathwayscited by 5
  10. 2022Recommender Systems and Algorithmic Hatecited by 5
  11. 2022The dangers in algorithms learning humans' values and irrationalitiescited by 3
  12. 2016PNP: Fast Path Ensemble Method for Movie Designcited by 2
  13. 2020Bandits Under The Influence (Extended Version)cited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 2; 2026: 3.

5 papers.

  1. 2025You Don't Bring Me Flowers: Mitigating Unwanted Recommendations Through Conformal Risk Controlcited by 3
  2. 2025Datasets for Navigating Sensitive Topics in Recommendation Systemscited by 2
  3. 2026With a Little Help From My Friends: Collective Manipulation in Risk-Controlling Recommender Systemscited by 2
  4. 2026Socially-Aware Recommender Systems Mitigate Opinion Clusterizationcited by 0
  5. 2026Gaming the Metric, Not the Harm: Certifying Safety Audits against Strategic Platform Manipulationcited by 0