arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.10322cs.LG
收藏

梯度屏蔽联邦优化

Gradient Masked Federated Optimization

MILA(MILA研究所);University of Montreal(蒙特利尔大学)

Irene Tenison, Sreya Francis, Irina Rish

摘要:

联邦平均(Federated Averaging,FedAVG)算法由于其简单性和低通信开销而成为最流行的联邦学习算法。我们使用简单的例子来说明FedAVG有将参与客户端的最优值缝合在一起的趋势。当在具有新数据分布的新客户端上使用时,这些缝合的优化表现出较差的泛化能力。受Arjovsky等人,2019; Parascandolo等人,2020年),我们专注于学习一个模型,该模型同时在不同的客户端上是局部最优的。我们提出了对FedAVG算法的修改以包括掩蔽梯度(来自Parascandolo等人,2020)),并使用它们来执行附加的服务器模型更新。我们表明,该算法实现了更好的准确性(外的分布)比FedAVG,特别是当数据是不相同的分布在客户端。

英文摘要:

Federated Averaging (FedAVG) has become the most popular federated learning algorithm due to its simplicity and low communication overhead. We use simple examples to show that FedAVG has the tendency to sew together the optima across the participating clients. These sewed optima exhibit poor generalization when used on a new client with new data distribution. Inspired by the invariance principles in (Arjovsky et al., 2019; Parascandolo et al., 2020), we focus on learning a model that is locally optimal across the different clients simultaneously. We propose a modification to FedAVG algorithm to include masked gradients (AND-mask from (Parascandolo et al., 2020)) across the clients and uses them to carry out an additional server model update. We show that this algorithm achieves better accuracy (out-of-distribution) than FedAVG, especially when the data is non-identically distributed across clients.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑