Distributed DP-Helmet: Scalable Differentially Private Non-interactive Averaging of Single Layers

1 code implementation • 3 Nov 2022 • Moritz Kirschte, Sebastian Meiser, Saman Ardalan, Esfandiar Mohammadi

We provide experimental evidence that blind averaging for SVMs and single Softmax-layer (Softmax-SLP) can have a strong utility-privacy tradeoff: we reach an accuracy of 86% on CIFAR-10 for $\varepsilon$ = 0. 4 and 1, 000 users, of 44% on CIFAR-100 for $\varepsilon$ = 1. 2 and 100 users, and of 39% on federated EMNIST for $\varepsilon$ = 0. 4 and 3, 400 users, all after a SimCLR-based pretraining.

Federated Learning