Graph-regularized federated learning with shareable side information

作者：

Highlights：

• Exploit the side information of clients for personalized federated learning.

• Propose a client-similarity graph regularized federated learning framework.

• Introduce to calculate the similarity between non-iid clients by data distributions.

• Conduct the evaluations on the personalization of federation frameworks.

摘要

•Exploit the side information of clients for personalized federated learning.•Propose a client-similarity graph regularized federated learning framework.•Introduce to calculate the similarity between non-iid clients by data distributions.•Conduct the evaluations on the personalization of federation frameworks.

论文关键词：Federated learning,Graph-regularized model,Similarity,Side information,Heterogeneous data classification

论文评审过程：Received 21 May 2022, Revised 27 September 2022, Accepted 27 September 2022, Available online 1 October 2022, Version of Record 17 October 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.109960