Personalized Cross-Silo Federated Learning on Non-IID Data
Abstract
Non-IID data present a tough challenge for federated learning. In this paper, we explore a novel idea of facilitating pairwise collaborations between clients with similar data. We propose FedAMP, a new method employing federated attentive message passing to facilitate similar clients to collaborate more. We establish the convergence of FedAMP for both convex and non-convex models, and propose a heuristic method to further improve the performance of FedAMP when clients adopt deep neural networks as personalized models. Our extensive experiments on benchmark data sets demonstrate the superior performance of the proposed methods.
Cite
@article{arxiv.2007.03797,
title = {Personalized Cross-Silo Federated Learning on Non-IID Data},
author = {Yutao Huang and Lingyang Chu and Zirui Zhou and Lanjun Wang and Jiangchuan Liu and Jian Pei and Yong Zhang},
journal= {arXiv preprint arXiv:2007.03797},
year = {2021}
}
Comments
Accepted by AAAI 2021. The API of this work is available at Huawei Cloud (https://developer.huaweicloud.com/develop/aigallery/notebook/detail?id=6d4a9521-6a4d-4b6d-b84d-943d7c7b1cbd), free registration is required before use