Privacy-Preserving Distributed Expectation Maximization for Gaussian Mixture Model Using Subspace Perturbation

Privacy has become a major concern in machine learning. In fact, the federated learning is motivated by the privacy concern as it does not allow to transmit the private data but only intermediate updates. However, federated learning does not always guarantee privacy-preservation as the intermediate...

Full description

Saved in:
Bibliographic Details
Published inICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 4263 - 4267
Main Authors Li, Qiongxiu, Gundersen, Jaron Skovsted, Tjell, Katrine, Wisniewski, Rafal, Christensen, Mads Grasboll
Format Conference Proceeding
LanguageEnglish
Published IEEE 23.05.2022
Subjects
Collaborative work
differential privacy
Federated learning
information-theoretic
Numerical models
Numerical simulation
Perturbation methods
Privacy
privacy-accuracy
secure multiparty computation
Signal processing
Signal processing algorithms
Online AccessGet full text

Cover

More Information
Summary:Privacy has become a major concern in machine learning. In fact, the federated learning is motivated by the privacy concern as it does not allow to transmit the private data but only intermediate updates. However, federated learning does not always guarantee privacy-preservation as the intermediate updates may also reveal sensitive information. In this paper, we give an explicit information-theoretical analysis of a federated expectation maximization algorithm for Gaussian mixture model and prove that the intermediate updates can cause severe privacy leakage. To address the privacy issue, we propose a fully decentralized privacy-preserving solution, which is able to securely compute the updates in each maximization step. Additionally, we consider two different types of security attacks: the honest-but-curious and eavesdropping adversary models. Numerical validation shows that the proposed approach has superior performance compared to the existing approach in terms of both the accuracy and privacy level.
ISSN:2379-190X
DOI:10.1109/ICASSP43922.2022.9746144