# Ranking with Popularity Bias: User Welfare under Self-Amplification Dynamics

arXiv: https://arxiv.org/abs/2305.18333 · HTML: https://mlsanity.com/paper/2305.18333/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 30 of its 85 references that are in the map.
- First posted on arXiv 24 May 2023.

## Builds on (most cited first, 30 of 30)

- [Sketching as a Tool for Numerical Linear Algebra](https://mlsanity.com/paper/1411.4357/) (2014)
- [Fairness of Exposure in Rankings](https://mlsanity.com/paper/1802.07281/) (2018)
- [Unbiased Learning-to-Rank with Biased Feedback](https://mlsanity.com/paper/1608.04468/) (2016)
- [Improved Optimistic Algorithms for Logistic Bandits](https://mlsanity.com/paper/2002.07530/) (2020)
- [Causal Intervention for Leveraging Popularity Bias in Recommendation](https://mlsanity.com/paper/2105.06067/) (2021)
- [Mostly Exploration-Free Algorithms for Contextual Bandits](https://mlsanity.com/paper/1704.09011/) (2017)
- [Managing Popularity Bias in Recommender Systems with Personalized Re-ranking](https://mlsanity.com/paper/1901.07555/) (2019)
- [Explore First, Exploit Next: The True Shape of Regret in Bandit Problems](https://mlsanity.com/paper/1602.07182/) (2016)
- [Cascading Bandits: Learning to Rank in the Cascade Model](https://mlsanity.com/paper/1502.02763/) (2015)
- [The Unfairness of Popularity Bias in Recommendation](https://mlsanity.com/paper/1907.13286/) (2019)
- [Controlling Fairness and Bias in Dynamic Learning-to-Rank](https://mlsanity.com/paper/2005.14713/) (2020)
- [Instance-Wise Minimax-Optimal Algorithms for Logistic Bandits](https://mlsanity.com/paper/2010.12642/) (2020)
- [Cascading Bandits for Large-Scale Recommendation Problems](https://mlsanity.com/paper/1603.05359/) (2016)
- [User-centered Evaluation of Popularity Bias in Recommender Systems](https://mlsanity.com/paper/2103.06364/) (2021)
- [The Connection Between Popularity Bias, Calibration, and Fairness in Recommendation](https://mlsanity.com/paper/2008.09273/) (2020)
- [Leveraging Good Representations in Linear Contextual Bandits](https://mlsanity.com/paper/2104.03781/) (2021)
- [OSOM: A simultaneously optimal algorithm for multi-armed and linear contextual bandits](https://mlsanity.com/paper/1905.10040/) (2019)
- [Online Learning to Rank in Stochastic Click Models](https://mlsanity.com/paper/1703.02527/) (2017)
- [Multi-sided Exposure Bias in Recommendation](https://mlsanity.com/paper/2006.15772/) (2020)
- [Popularity Bias Is Not Always Evil: Disentangling Benign and Harmful Bias for Recommendation](https://mlsanity.com/paper/2109.07946/) (2021)
- [Bandits with Partially Observable Confounded Data](https://mlsanity.com/paper/2006.06731/) (2020)
- [Semiparametric Contextual Bandits](https://mlsanity.com/paper/1803.04204/) (2018)
- [Thompson Sampling Algorithms for Cascading Bandits](https://mlsanity.com/paper/1810.01187/) (2018)
- [UCB-based Algorithms for Multinomial Logistic Regression Bandits](https://mlsanity.com/paper/2103.11489/) (2021)
- [Reinforcement Learning with History-Dependent Dynamic Contexts](https://mlsanity.com/paper/2302.02061/) (2023)
- [Popularity Bias in Recommendation: A Multi-stakeholder Perspective](https://mlsanity.com/paper/2008.08551/) (2020)
- [Reinforcement Learning with a Terminator](https://mlsanity.com/paper/2205.15376/) (2022)
- [Disentangling User Interest and Conformity for Recommendation with Causal Embedding](https://mlsanity.com/paper/2006.11011/) (2020)
- [Evolution of Popularity Bias: Empirical Study and Debiasing](https://mlsanity.com/paper/2207.03372/) (2022)
- [Overcoming Prior Misspecification in Online Learning to Rank](https://mlsanity.com/paper/2301.10651/) (2023)

## Abstract (arXiv)

While popularity bias is recognized to play a crucial role in recommmender (and other ranking-based) systems, detailed analysis of its impact on collective user welfare has largely been lacking. We propose and theoretically analyze a general mechanism, rooted in many of the models proposed in the literature, by which item popularity, item quality, and position bias jointly impact user choice. We focus on a standard setting in which user utility is largely driven by item quality, and a recommender attempts to estimate it given user behavior. Formulating the problem as a non-stationary contextual bandit, we study the ability of a recommender policy to maximize user welfare under this model. We highlight the importance of exploration, not to eliminate popularity bias, but to mitigate its negative impact on welfare. We first show that naive popularity-biased recommenders induce linear regret by conflating item quality and popularity. More generally, we show that, even in linear settings, identifiability of item quality may not be possible due to the confounding effects of popularity bias. However, under sufficient variability assumptions, we develop an efficient optimistic algorithm and prove efficient regret guarantees w.r.t. user welfare. We complement our analysis with several simulation studies, which demonstrate the negative impact of popularity bias on the performance of several natural recommender policies.