LotteryFL: Personalized and Communication-Efficient Federated Learning with Lottery Ticket Hypothesis on Non-IID Datasets阅读笔记
针对FL的问题:
1. statistical heterogeneity
2. communication efficiency
主要优化策略:
使用Lottery Ticket hypothesis,因为worker上面只是子网络,所以通讯量下降了。因为worker获取的就是特异化的子网络,所以不需要fine-tunes,节约了personalized的时间。
论文贡献点:
- 提出Lottery FL,实现了personalization and high communication efficiency under non-IID setting。
- 建立一个新metric:Client-Wise Non-IID Index,目的:evaluate the degree of non-IID data distribution across clients(没细看)
- 比较了FedAvg和LG-FedAvg
下一篇论文:LG-FedAvg
关于 FedAvg:blog链接