ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2005.00783
15
2
v1v2 (latest)

Differentially Private Generation of Small Images

2 May 2020
J. Schwabedal
Pascal Michel
Mario S. Riontino
ArXiv (abs)PDFHTML
Abstract

We explore the training of generative adversarial networks with differential privacy to anonymize image data sets. On MNIST, we numerically measure the privacy-utility trade-off using parameters from ϵ\epsilonϵ-δ\deltaδ differential privacy and the inception score. Our experiments uncover a saturated training regime where an increasing privacy budget adds little to the quality of generated images. We also explain analytically why differentially private Adam optimization is independent of the gradient clipping parameter. Furthermore, we highlight common errors in previous works on differentially private deep learning, which we uncovered in recent literature. Throughout the treatment of the subject, we hope to prevent erroneous estimates of anonymity in the future.

View on arXiv
Comments on this paper