Renyi Differential Privacy

2017/02/28 by Ilya Mironov · 6 citations
Computer Science · #cs.CR

paper · pdf · doi:10.1109/csf.2017.11

published as Proceedings of IEEE 30th Computer Security Foundations Symposium CSF 2017, pages 263-275, IEEE, Aug 2017

arxiv created 2017/08/25 · arxiv updated 2018/12/12

Abstract

We propose a natural relaxation of differential privacy based on the Renyi divergence. Closely related notions have appeared in several recent papers that analyzed composition of differentially private mechanisms. We argue that the useful analytical tool can be used as a privacy definition, compactly and accurately representing guarantees on the tails of the privacy loss. We demonstrate that the new definition shares many important properties with the standard definition of differential privacy, while additionally allowing tighter analysis of composite heterogeneous mechanisms.

Cited by

• Privacy-Preserving Federated Autoencoder for ECG Anomaly Detection on Edge Devices
• Differential privacy representation geometry for medical image analysis
• PriDyG: Privacy-preserving Dynamic Graph Inference with LLM-GNN Collaboration
• MCMC for Bayesian estimation of Differential Privacy from Membership Inference Attacks
• A Comprehensive Survey of Synthetic Tabular Data Generation
• Benchmarking Differentially Private Tabular Data Synthesis

Related

• Concentrated Differential Privacy: Simplifications, Extensions, and Lower Bounds
• A Better Bound Gives a Hundred Rounds: Enhanced Privacy Guarantees via f-Divergences
• On Rényi Differential Privacy in Statistics-Based Synthetic Data Generation
• Lower Bounds for Rényi Differential Privacy in a Black-Box Setting
• Total Variation Meets Differential Privacy