Mitigating Popularity Bias in Recommendation with Unbalanced Interactions: A Gradient Perspective

no code implementations • 31 Oct 2022 • Weijieying Ren, Lei Wang, Kunpeng Liu, Ruocheng Guo, Lim Ee Peng, Yanjie Fu

We present a gradient perspective to understand two negative impacts of popularity bias in recommendation model optimization: (i) the gradient direction of popular item embeddings is closer to that of positive interactions, and (ii) the magnitude of positive gradient for popular items are much greater than that of unpopular items.

Model Optimization Recommendation Systems