Rényi Differential Privacy에서 Group Privacy Amplification 및 Unified Amplification by Subsampling
강화를 통해 그룹 프라이버시를 강화하고, 샘플링을 통한 통합된 증폭을 통해 Rényi Differential Privacy를 다룹니다.
Private, Efficient, and Optimal K-Norm and Elliptic Gaussian Noise for Sum, Count, and Vote
Efficiently implementing optimal K-norm mechanisms for Sum, Count, and Vote with Elliptic Gaussian Noise.
Differential Privacy Lower Bounds for Continual Observation and Online Threshold Queries
Lower bounds for differential privacy under continual observation and online threshold queries are crucial for understanding the price of privacy over time.
Analyzing Privacy in DP-SGD Mechanisms
There can be significant differences in privacy guarantees between different batch sampling methods in DP-SGD mechanisms.
Analyzing Privacy of Selection Mechanisms with Gaussian Noise
Pure-DP bounds for Gaussian selection mechanisms with bounded queries.
SoK: Trajectory Generation for Privacy and Utility
Synthetic trajectory data generation offers privacy while maintaining utility, but current solutions lack formal privacy guarantees and face limitations in practical evaluations.
Group Privacy Amplification and Unified Amplification by Subsampling for Rényi Differential Privacy
The author explores the joint study of group privacy and amplification by subsampling in the context of Rényi-DP, providing a unified framework for deriving guarantees. This approach leads to novel and tight guarantees for various mechanisms.
Achieving Pure Differential Privacy for Functional Summaries via Independent Component Laplace Process
The authors introduce the Independent Component Laplace Process (ICLP) mechanism to achieve pure differential privacy for functional summaries, addressing limitations of existing mechanisms by treating summaries as truly infinite-dimensional objects.
© 2024 by Linnk AI