Mitigating Popularity Bias in Recommendation with Global Listwise Learning and Progressive Bi-Weighting

2026-09-28

AI Summary

This study addresses the popularity bias induced by long-tailed distributions in recommender systems, as well as the limitations of existing Inverse Propensity Scoring (IPS) methods, which are only locally unbiased and suffer from inaccurate propensity estimation. To overcome these challenges, this work proposes Mult-BiW, a globally debiased framework that constructs a globally unbiased objective based on multinomial likelihood and inverse propensity scores. It introduces a novel dual-weighting strategy that jointly models propensity and collection, along with a smoothing mechanism to refine propensity estimation. Furthermore, a progressive weighting scheme is incorporated to effectively mitigate the adverse effects of aggressive reweighting on representation learning. Experiments on real-world datasets demonstrate that the proposed framework consistently outperforms state-of-the-art baseline models.

Abstract

In recommender systems, user feedback typically follows a long-tail distribution, which leads many recommendation algorithms to exacerbate popularity bias by disproportionately favoring popular items. To mitigate this issue, recent studies have employed Inverse Propensity Scoring (IPS) to rebalance training data via reweighting user-item interactions. However, the effectiveness of IPS-based approaches is often constrained by locally unbiased objectives and inaccurate propensity estimation. In this paper, we propose Multinomial Likelihood with Bi-Weighting (Mult-BiW) to address these limitations. First, we introduce a debiasing framework, termed Mult-IPS, which integrates multinomial likelihood with IPS to capture global and unbiased user preferences over the entire item set. Second, we develop a Bi-Weighting (BiW) strategy that jointly leverages propensity scores and a collection model, incorporating a smoothing mechanism to enhance the robustness of propensity estimation. We further provide theoretical analyses that establish an upper bound on the empirical bias and characterize the optimal form of the collection model. Third, to mitigate the adverse effects of aggressive reweighting on representation learning, we design a Progressive Bi-Weighting strategy that gradually transitions from discriminative representation learning to popularity debiasing. Extensive experiments on real-world datasets show that Mult-BiW consistently outperforms state-of-the-art baselines.

Problem

• Popularity Bias
• Recommender Systems
• Inverse Propensity Scoring
• Debiasing
• Long-tail Distribution

Innovation

• Popularity Bias
• Inverse Propensity Scoring
• Global Listwise Learning
• Bi-Weighting
• Progressive Debiasing