Geometry of Sensitivity: Twice Sampling and Hybrid Clipping in Differential Privacy with Optimal Gaussian Noise and Application to Deep Learning
We study the fundamental problem of the construction of optimal randomization in Differential Privacy (DP). Depending on the clipping strategy or additional properties of the processing function, the corresponding sensitivity set theoretically determines the necessary randomization to produce the re...
Main Authors: | Xiao, Hanshen, Wan, Jun, Devadas, Srinivas |
---|---|
Other Authors: | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
Format: | Article |
Language: | English |
Published: |
ACM|Proceedings of the 2023 ACM SIGSAC Conference on Computer and Communications Security
2023
|
Online Access: | https://hdl.handle.net/1721.1/153139 |
Similar Items
-
Local differential privacy in decentralized optimization
by: Xiao, Hanshen.
Published: (2020) -
Formal Privacy Proof of Data Encoding: The Possibility and Impossibility of Learnable Encryption
by: Xiao, Hanshen, et al.
Published: (2025) -
Designing education for twice-exceptional learners
by: Stillman, John Francis
Published: (2018) -
You never step in the same river twice.
by: Choo, Yan Bing.
Published: (2011) -
Sample-Efficient Proper PAC Learning with Approximate Differential Privacy
by: Ghazi, Badih, et al.
Published: (2022)