Federated optimization via knowledge codistillation

作者：

Highlights：

• A federated optimization framework based on knowledge codistillation is proposed.

• An extension is presented to hold a personalized model for each federated device.

• Theoretical convergence guarantees for our algorithms are provided.

• Performance of the proposed schemes is evaluated on diverse federated benchmarks.

摘要

•A federated optimization framework based on knowledge codistillation is proposed.•An extension is presented to hold a personalized model for each federated device.•Theoretical convergence guarantees for our algorithms are provided.•Performance of the proposed schemes is evaluated on diverse federated benchmarks.

论文关键词：Federated learning,Distributed computing,Federated optimization,Knowledge distillation,Non-IID data

论文评审过程：Received 17 April 2021, Revised 23 November 2021, Accepted 25 November 2021, Available online 11 December 2021, Version of Record 21 December 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.116310