mlsanityData to 7 Oct 2026

The Saddle-Point Accountant for Differential Privacy

arXiv 2208.09595 ↗First posted 20 Aug 2022cs.CR cs.IT cs.LGcited by 5

Not on the map. Why?

Abstract (arXiv)

We introduce a new differential privacy (DP) accountant called the saddle-point accountant (SPA). SPA approximates privacy guarantees for the composition of DP mechanisms in an accurate and fast manner. Our approach is inspired by the saddle-point method -- a ubiquitous numerical technique in statistics. We prove rigorous performance guarantees by deriving upper and lower bounds for the approximation error offered by SPA. The crux of SPA is a combination of large-deviation methods with central limit theorems, which we derive via exponentially tilting the privacy loss random variables corresponding to the DP mechanisms. One key advantage of SPA is that it runs in constant time for the n-fold composition of a privacy mechanism. Numerical experiments demonstrate that SPA achieves comparable accuracy to state-of-the-art accounting methods with a faster runtime.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 17 of its 43 references that are in the map. First posted on arXiv 20 Aug 2022.

Builds on

17 of 43 references.

References · 17

  1. 2016Deep Learning with Differential Privacycited by 2,682
  2. 2017Renyi Differential Privacycited by 643
  3. 2013The Composition Theorem for Differential Privacycited by 290
  4. 2018Privacy Amplification by Subsampling: Tight Analyses via Couplings and Divergencescited by 237
  5. 2018Improving the Gaussian Mechanism for Differential Privacy: Analytical Calibration and Optimal Denoisingcited by 216
  6. 2019R'enyi Differential Privacy of the Sampled Gaussian Mechanismcited by 199
  7. 2021Numerical Composition of Differential Privacycited by 174
  8. 2020The Discrete Gaussian for Differential Privacycited by 134
  9. 2019Hypothesis Testing Interpretations and Renyi Differential Privacycited by 84
  10. 2021Optimal Accounting of Differential Privacy via Characteristic Functioncited by 83
  11. 2022Connect the Dots: Tighter Discrete Approximations of Privacy Loss Distributionscited by 50
  12. 2020Tight Differential Privacy for Discrete-Valued Mechanisms and for the Subsampled Gaussian Mechanism Using FFTcited by 43
5 more
  1. 2020A Better Bound Gives a Hundred Rounds: Enhanced Privacy Guarantees via f-Divergencescited by 21
  2. 2021Computing Differential Privacy Guarantees for Heterogeneous Compositions Using FFTcited by 20
  3. 2017R'enyi Differential Privacy Mechanisms for Posterior Samplingcited by 15
  4. 2022Edgeworth Accountant: An Analytical Approach to Differential Privacy Compositioncited by 14
  5. 2022Faster Privacy Accounting via Evolving Discretizationcited by 14

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 2; 2024: 1; 2025: 1.

5 papers.

  1. 2022Composition of Differential Privacy & Privacy Amplification by Subsamplingcited by 42
  2. 2023Bounding data reconstruction attacks with the hypothesis testing interpretation of differential privacycited by 14
  3. 2023A Randomized Approach for Tight Privacy Accountingcited by 11
  4. 2025Unifying Re-Identification, Attribute Inference, and Data Reconstruction Risks in Differential Privacycited by 11
  5. 2024Attack-Aware Noise Calibration for Differential Privacycited by 6