Reproduce “Mitigating Popularity Bias in Recommendation with Global Listwise Learning and Progressive Bi-Weighting” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Tianyu Zhu, Jiandong Ding, Yansong Shi, Guoqing Chen, Jian-Yun Nie
Mult-BiW addresses popularity skew in recommendation by training users’ relevant items to compete across the catalog, while smoothing estimated exposure propensities and applying individualized weights progressively. On six datasets, it improves average Recall@20 by 9.43% and NDCG@20 by 14.80% over each dataset’s strongest baseline. These are offline results, not evidence of universal metric gains or improved real-world outcomes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.