BriefGPT.xyz
May, 2019
自适应剪裁差分隐私学习
Differentially Private Learning with Adaptive Clipping
HTML
PDF
Om Thakkar, Galen Andrew, H. Brendan McMahan
TL;DR
该论文提出了一种新的针对联合学习中用户级差分隐私训练神经网络的方法,它通过自适应裁剪来解决不同设备以及不同学习任务下的裁剪行为的设置问题,并通过实验表明该方法与现有的裁剪方式相比具有显著的性能提升。
Abstract
We introduce a new
adaptive clipping
technique for training learning models with user-level
differential privacy
that removes the need for extensive parameter tuning. Previous approaches to this problem use the F
→