Navigating Alignment for Non-identical Client Class Sets: A Label Name-Anchored Federated Learning Framework. KDD 2023. [arXiv]
Check our 2-min promotional video on YouTube!
Download and put the datasets in data folder. The datasets we used are listed below.
ExtraSensory: http://extrasensory.ucsd.edu/
MIMIC-III: https://physionet.org/content/mimiciii/1.4/. Please follow this GitHub repo to preprocess the data.
PAMAP2: https://archive.ics.uci.edu/dataset/231/pamap2+physical+activity+monitoring. Please refer to the code for data preprocessing.
Reuters-21578 R8: https://ana.cachopo.org/datasets-for-single-label-text-categorization
Below are the packages we used for our experiments.
python==3.9.16
networkx==2.8.8
node2vec==0.4.6
numpy==1.21.4
pandas==1.5.3
pytorch_nlp==0.5.0
scikit_learn==1.2.2
torch==1.11.0
tqdm==4.65.0
Specify the task by -t and the gpu device by -g. For example, to run FedAlign on PAMAP2 with gpu 6, run:
python main.py -t pamap2 --fedalign -g 6
Please cite the following paper if you found our framework useful. Thanks!
@inproceedings{zhang2023navigating,
title={Navigating Alignment for Non-identical Client Class Sets: A Label Name-Anchored Federated Learning Framework},
author={Zhang, Jiayun and Zhang, Xiyuan and Zhang, Xinyang and Hong, Dezhi and Gupta, Rajesh K and Shang, Jingbo},
booktitle={Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
pages={3297--3308},
year={2023}
}