alpemreacar / PFL

Debiasing Model Updates for Improving Personalized Federated Training

This is implementation of Debiasing Model Updates for Improving Personalized Federated Training.

Requirements

Please install the required packages. The code is compiled with Python 3.7 dependencies in a virtual environment via

pip install -r requirements.txt

Instructions

An example code for CIFAr-10, ACID, 5 class per device setting is given. Run

python cifar10_ACID.py

The code,

• Constructs a federated dataset,

• Trains all methods,

• Plots the average test accuracy vs. rounds convergence curves.

Citation

@InProceedings{pmlr-v139-acar21a,
  title = {Debiasing Model Updates for Improving Personalized Federated Training},
  author = {Acar, Durmus Alp Emre and Zhao, Yue and Zhu, Ruizhao and Matas, Ramon and Mattina, Matthew and Whatmough, Paul and Saligrama, Venkatesh},
  booktitle = {Proceedings of the 38th International Conference on Machine Learning},
  pages = {21--31},
  year = {2021},
  editor = {Meila, Marina and Zhang, Tong},
  volume = {139},
  series = {Proceedings of Machine Learning Research},
  month = {18--24 Jul},
  publisher = {PMLR},
  pdf =  {http://proceedings.mlr.press/v139/acar21a/acar21a.pdf},
  url =  {http://proceedings.mlr.press/v139/acar21a.html}
}