Differentially Private Generative Adversarial Networks with Model Inversion

Dongjie Chen, Sen Ching Samson Cheung, Chen Nee Chuah, Sally Ozonoff

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

To protect sensitive data in training a Generative Adversarial Network (GAN), the standard approach is to use differentially private (DP) stochastic gradient descent method in which controlled noise is added to the gradients. The quality of the output synthetic samples can be adversely affected and the training of the network may not even converge in the presence of these noises. We propose Differentially Private Model Inversion (DPMI) method where the private data is first mapped to the latent space via a public generator, followed by a lower-dimensional DP-GAN with better convergent properties. Experimental results on standard datasets CIFAR10 and SVHN as well as on a facial landmark dataset for Autism screening show that our approach outperforms the standard DP-GAN method based on Inception Score, Frechet Inception Distance, and classification accuracy under the same privacy guarantee.

Original languageEnglish (US)
Title of host publication2021 IEEE International Workshop on Information Forensics and Security, WIFS 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665417174
DOIs
StatePublished - 2021
Event2021 IEEE International Workshop on Information Forensics and Security, WIFS 2021 - Montpellier, France
Duration: Dec 7 2021Dec 10 2021

Publication series

Name2021 IEEE International Workshop on Information Forensics and Security, WIFS 2021

Conference

Conference2021 IEEE International Workshop on Information Forensics and Security, WIFS 2021
Country/TerritoryFrance
CityMontpellier
Period12/7/2112/10/21

Keywords

  • differential privacy
  • Generative adversarial networks
  • model inversion

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Information Systems and Management
  • Safety, Risk, Reliability and Quality

Fingerprint

Dive into the research topics of 'Differentially Private Generative Adversarial Networks with Model Inversion'. Together they form a unique fingerprint.

Cite this