The Distributed Discrete Gaussian Mechanism for Federated Learning with Secure Aggregation

02/12/2021
by   Peter Kairouz, et al.
0

We consider training models on private data that is distributed across user devices. To ensure privacy, we add on-device noise and use secure aggregation so that only the noisy sum is revealed to the server. We present a comprehensive end-to-end system, which appropriately discretizes the data and adds discrete Gaussian noise before performing secure aggregation. We provide a novel privacy analysis for sums of discrete Gaussians. We also analyze the effect of rounding the input data and the modular summation arithmetic. Our theoretical guarantees highlight the complex tension between communication, privacy, and accuracy. Our extensive experimental results demonstrate that our solution is essentially able to achieve a comparable accuracy to central differential privacy with 16 bits of precision per value.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
10/11/2021

The Skellam Mechanism for Differentially Private Federated Learning

We introduce the multi-dimensional Skellam mechanism, a discrete differe...
research
07/09/2022

The Poisson binomial mechanism for secure and private federated learning

We introduce the Poisson Binomial mechanism (PBM), a discrete differenti...
research
11/18/2022

The communication cost of security and privacy in federated frequency estimation

We consider the federated frequency estimation problem, where each user ...
research
03/07/2022

The Fundamental Price of Secure Aggregation in Differentially Private Federated Learning

We consider the problem of training a d dimensional model with distribut...
research
11/30/2019

Federated Learning with Autotuned Communication-Efficient Secure Aggregation

Federated Learning enables mobile devices to collaboratively learn a sha...
research
10/22/2022

Federated Calibration and Evaluation of Binary Classifiers

We address two major obstacles to practical use of supervised classifier...
research
06/08/2021

Private Multi-Group Aggregation

We study the differentially private multi group aggregation (PMGA) probl...

Please sign up or login with your details

Forgot password? Click here to reset