When Recommendation Denoising Meets Popularity Bias: Understanding and Mitigating Their Interaction

2026-06-11

AI Summary

This work addresses a critical yet overlooked issue in implicit-feedback recommendation systems: existing small-loss denoising methods often misidentify genuine interactions with long-tail items as noise, thereby exacerbating popularity bias. The study is the first to uncover the coupling between denoising mechanisms and popularity bias, introducing a conditional redistribution theory. Building on this insight, the authors propose Popularity-Aware Denoising (PAD), a lightweight plug-in strategy that dynamically reweights small-loss samples by analyzing loss distributions, modeling the effective signal ratio between head and tail items, and leveraging popularity-stratified item groups. PAD is compatible with diverse backbone models and consistently outperforms state-of-the-art denoising approaches across three datasets and three architectures, notably achieving a superior balance between accuracy and diversity in matrix factorization models.

Abstract

Implicit feedback is the dominant data source for recommender systems, but behavioral logs are often contaminated by false-positive interactions caused by mis-clicks, biased exposure, and interface effects. Denoising recommendation methods improve robustness by down-weighting or filtering interactions suspected to be noisy, often relying on the small-loss heuristic. We revisit this heuristic through the lens of popularity bias. Tail-item positives can be harder to fit because they are sparsely observed, and thus may receive larger losses even when they reflect genuine user preference. Under such popularity-dependent loss patterns, monotone loss-based reweighting can suppress clean-but-hard tail signals and increase the head-tail imbalance in effective supervision. We formalize this interaction through the effective head-tail signal ratio induced by denoising weights and derive a conditional reallocation result: when the loss distribution of tail positives is right-shifted relative to that of head positives, small-loss reweighting increases the effective head-tail signal ratio compared with ERM. Motivated by this analysis, we propose Popularity-Aware Denoising (PAD), a lightweight plug-in framework that modulates denoising strength by item popularity. PAD applies stronger denoising to highly exposed items while being more conservative on tail items, preserving more clean-but-hard long-tail signals. Experiments on three datasets and three backbones show that PAD generally improves over representative denoising baselines and provides favorable accuracy-diversity tradeoffs, especially on MF-style recommenders.

Problem

• recommendation denoising
• popularity bias
• implicit feedback
• head-tail imbalance
• false-positive interactions

Innovation

• popularity bias
• denoising recommendation
• small-loss heuristic
• long-tail items
• signal reallocation