FedNoRo

This is 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.

Brief Introduction

This paper proposes a federated noisy label learning framework for class-imbalanced and heterogeneous multi-source medical data.

Dataset

Please download the ICH dataset from kaggle and preprocess it follow this notebook. Please download the ISIC 2019 dataset from this link. Data partition can be found in the paper.

Update (Mar. 2024): You may get the ICH dataset here.

Requirements

We recommend using conda to setup the environment. See the requirements.txt for environment configuration.

Main Baselines:

• FedAvg [paper]
• FedProx [paper]
• FedLA (Logit Adjustment) [paper]
• RoFL [paper] [code]
• RHFL [paper] [code]
• FedLSR [paper] [code]
• FedCorr [paper] [code]

Citation

If this repository is useful for your research, please consider citing:

@inproceedings{wu2023fednoro,
  title     = {FedNoRo: Towards Noise-Robust Federated Learning by Addressing Class Imbalance and Label Noise Heterogeneity},
  author    = {Wu, Nannan and Yu, Li and Jiang, Xuefeng and Cheng, Kwang-Ting and Yan, Zengqiang},
  booktitle = {Proceedings of the Thirty-Second International Joint Conference on
               Artificial Intelligence, {IJCAI-23}},
  pages     = {4424--4432},
  year      = {2023},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2023/492},
  url       = {https://doi.org/10.24963/ijcai.2023/492},
}

Contact

For any questions, please contact '[email protected]'.