FedNoRo: Towards Noise-Robust Federated Learning by Addressing Class Imbalance and Label Noise Heterogeneity

Nannan Wu, Li Yu, Xuefeng Jiang, Kwang-Ting Cheng, Zengqiang Yan
5/9/2023
cs.LGcs.AI

No official code found. Best implementation: wnn2000/FedNoRo (29 stars, last commit 2024-03-16).

Abstract

Federated noisy label learning (FNLL) is emerging as a promising tool for privacy-preserving multi-source decentralized learning. Existing research, relying on the assumption of class-balanced global data, might be incapable to model complicated label noise, especially in medical scenarios. In this paper, we first formulate a new and more realistic federated label noise problem where global data is class-imbalanced and label noise is heterogeneous, and then propose a two-stage framework named FedNoRo for noise-robust federated learning. Specifically, in the first stage of FedNoRo, per-class loss indicators followed by Gaussian Mixture Model are deployed for noisy client identification. In the second stage, knowledge distillation and a distance-aware aggregation function are jointly adopted for noise-robust federated model updating. Experimental results on the widely-used ICH and ISIC2019 datasets demonstrate the superiority of FedNoRo against the state-of-the-art FNLL methods for addressing class imbalance and label noise heterogeneity in real-world FL scenarios.

View on arXivView PDF

Code Implementations(1)

wnn2000/FedNoRoCommunity100%

The official PyTorch implementation for the paper: "FedNoRo: Towards Noise-Robust Federated Learning By Addressing Class Imbalance and Label Noise Heterogeneity", which is accepted at IJCAI'23 main track.

2911PythonApr 21, 20232 years ago

Similar Papers

Learning Cautiously in Federated Learning with Noisy and Heterogeneous Clients

91% match2023

FedFixer: Mitigating Heterogeneous Label Noise in Federated Learning

89% match2024

Tackling Noisy Clients in Federated Learning with End-to-end Label Correction

89% match1 repo242024

FedNoisy: Federated Noisy Label Learning Benchmark

88% match1 repo292023

FedGSCA: Medical Federated Learning with Global Sample Selector and Client Adaptive Adjuster under Label Noise

88% match2025

Collaboratively Learning Federated Models from Noisy Decentralized Data

87% match1 repo12024

Cite this paper

@article{wu2023fednoro,
  title  = {FedNoRo: Towards Noise-Robust Federated Learning by Addressing Class Imbalance and Label Noise Heterogeneity},
  author = {Nannan Wu and Li Yu and Xuefeng Jiang and Kwang-Ting Cheng and Zengqiang Yan},
  year   = {2023},
  eprint = {2305.05230},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.05230v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.05230.svg)](https://codeofpaper.com/paper/2305.05230)

Discussion