提出全局修正器，让联邦学习在噪声标签下仍能保持稳定性能
Learning Locally, Revising Globally: Global Reviser for Federated Learning with Noisy Labels
- 利用全局模型缓慢记忆真实标签的特性，设计三模块协同修正噪声标签
- 在三个主流噪声标签基准上超越8个前沿方法，极端异构环境下仍有效
- 适合实际场景中标签质量差、数据分布不均的联邦学习应用
传统联邦学习高度依赖高质量标签，但现实中常面临标签噪声问题（F-LN），且因客户端数据分布异构，噪声类型与比例各异，使问题更严重。本研究首次观察到：联邦学习的全局模型会缓慢记忆真实标签，具备维持可靠预测与鲁棒表示的能力。受此启发，提出联邦全局修正器（FedGR），一种简单有效的三模块协同方法，通过利用该内在特性，在无需外部监督的情况下自洽地修正噪声标签并正则化本地训练。在三个广泛使用的F-LN基准上的大量实验表明，FedGR持续优于八个先进基线方法，即使在严重标签噪声和数据异构条件下也表现卓越。
原文摘要 · Abstract (English)
Conventional federated learning (FL) heavily depends on high-quality labels, which are often impractical in the real world, leading to the federated label-noise (F-LN) problem. Worse still, the F-LN problem is exacerbated by the heterogeneity of FL, whereas clients experience different label-noise types, ratios, and data distribution. In this study, we first observe an intriguing phenomenon that the global model of FL exhibits a slow memorization of noisy labels, suggesting its ability to maintain reliable predictions and robust representations in FL. Motivated by this, we propose a novel method termed Federated Global Reviser (\method), a straightforward yet effective method comprising three modules that collaboratively rectify noisy labels and regularize local training. By exploiting this inherent property, \method\ improves the label-noise robustness of FL in a self-contained manner. Extensive experiments on three widely used F-LN benchmarks demonstrate the superior performance of FedGR, consistently outperforming eight state-of-the-art baselines even in severe label-noise and data heterogeneity. Code: https://github.com/cs-yuxintian/FedGR-ICML26
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。