cpSGD: Communication-efficient and differentially-private distributed SGD

Abstract

Distributed stochastic gradient descent is an important subroutine in distributed learning. A setting of particular interest is when the clients are mobile devices, where two important concerns are communication efficiency and the privacy of the clients. Several recent works have focused on reducing the communication cost or introducing privacy guarantees, but none of the proposed communication efficient methods are known to be privacy preserving and none of the known privacy mechanisms are known to be communication efficient. To this end, we study algorithms that achieve both communication efficiency and differential privacy. For $d$ variables and $n\approx d$ clients, the proposed method uses $O(\log \log (nd))$ bits of communication per client per coordinate and ensures constant privacy. We also extend and improve previous analysis of the \emph{Binomial mechanism} showing that it achieves nearly the same utility as the Gaussian mechanism, while requiring fewer representation bits, which can be of independent interest.

References in corpus (1)

• QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding

Cited by in corpus (8)

• Federated Learning for Healthcare Informatics
• A review of distributed statistical inference
• DataLens: Scalable Privacy Preserving Training via Gradient Compression and Aggregation
• Group privacy for personalized federated learning
• Coded Computing for Master-Aided Distributed Computing Systems
• Decentralized Differentially Private Without-Replacement Stochastic Gradient Descent
• An Efficient DP-SGD Mechanism for Large Scale NLP Models
• FED-${\chi }^2$: Privacy Preserving Federated Correlation Test