Research map / Federated learning and privacy / Differential privacy
Improving the Gaussian Mechanism for Differential Privacy: Analytical Calibration and Optimal Denoising
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace classical Gaussian tail-bound calibration with numerical calibration using the Gaussian CDF | Gaussian mechanism · Gaussian noise calibration | noise variance: 1.4 | Helped |
| Use analytic calibration instead of classical calibration | Gaussian mechanism · Gaussian noise calibration | noise variance | Helped |
| Add James-Stein denoising | Analytic Gaussian mechanism · Private mean estimation | L2 error | Helped |
| Add soft-thresholding denoising | Analytic Gaussian mechanism · Private mean estimation | L2 error | Helped |
| Use James-Stein denoising rather than soft-thresholding | Analytic Gaussian mechanism for mean estimation · High-dimensional private mean estimation | L2 error | Helped |
| Add denoising to the analytic Gaussian mechanism | Analytic Gaussian mechanism · Private histogram release | L1 error | Helped |
| Use Laplace noise rather than analytic Gaussian noise | Private histogram release · ε = 1 | L1 error | Helped |
| Use analytic Gaussian noise rather than Laplace noise | Private histogram release · Lower ε values | L1 error | Helped |
| Use soft-thresholding rather than James-Stein shrinking | Analytic Gaussian mechanism for histogram release · Sparse private histogram release | L1 error | Helped |
| Apply graph trend filtering rather than wavelet smoothing | Analytic Gaussian mechanism for NYC Taxi heat maps · NYC Taxi heat maps | MSE | Helped |
| Apply denoising as post-processing | Analytic Gaussian mechanism for NYC Taxi heat maps · NYC Taxi heat maps | MSE | Helped |
| Use analytic calibration rather than classical calibration | Gaussian mechanism for NYC Taxi heat maps · NYC Taxi heat maps | MSE | Helped |
Closest papers by meaning
- Private Adaptive Covariance Estimation via Gaussian Graphical Models (NeurIPS 2026 · similarity 0.90)
- Privacy Guarantees in Posterior Sampling under Contamination (NeurIPS 2026 · similarity 0.90)
- The Gaussian Mixing Mechanism: Renyi Differential Privacy via Gaussian Sketches (NeurIPS 2025 · similarity 0.89)
- Improved Communication-Privacy Trade-offs in $L_2$ Mean Estimation under Streaming Differential Privacy (ICML 2024 · similarity 0.89)
- Optimizing Noise Distributions for Differential Privacy (ICML 2025 · similarity 0.89)
- Instance-Specific Asymmetric Sensitivity in Differential Privacy (NeurIPS 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.