Research map / Federated learning and privacy / Differential privacy
Instance-optimal Mean Estimation Under Differential Privacy
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 |
|---|---|---|---|
| use shifted-clipped-mean mechanism instead of COINPRESS MVMRec with best t | DP mean estimation · Gaussian N(mu,Sigma) statistical mean estimation | l2 error | Helped |
| use shifted-clipped-mean instead of MVMRec when covariance is non-identity | DP mean estimation · Gaussian N(0,Sigma(kappa)) | l2 error | Helped |
| use shifted-clipped-mean instead of COINPRESS | DP empirical mean estimation · MNIST (d=784) | l2 error | Helped |
| estimate covariance first (MVCRec-MVMRec) instead of MVMRec alone | COINPRESS · Gaussian, varying d | l2 error | Mixed |
| remove the shift (centering) step from the clipped-mean estimator (ablation) | Shifted-CM · Gaussian with varying mean/R | l2 error | Hurt |
| use clipping quantile n - sqrt(2d/rho) instead of fixed quantiles (ablation) | clipped-mean estimator · synthetic dataset, n=500 | l2 error | Helped |
| use proposed LDP mean estimator instead of KnownVar coordinate-wise baseline | local-model mean estimation · Gaussian, n=1e5 | l2 error | Helped |
Closest papers by meaning
- Mean Estimation in the Add-Remove Model of Differential Privacy (ICML 2024 · similarity 0.90)
- Optimal Survey Design for Private Mean Estimation (ICML 2025 · similarity 0.89)
- Instance-Specific Asymmetric Sensitivity in Differential Privacy (NeurIPS 2024 · similarity 0.88)
- Optimal Differentially Private Model Training with Public Data (ICML 2024 · similarity 0.88)
- A Huber Loss Minimization Approach to Mean Estimation under User-level Differential Privacy (NeurIPS 2024 · similarity 0.88)
- Improved Communication-Privacy Trade-offs in $L_2$ Mean Estimation under Streaming Differential Privacy (ICML 2024 · similarity 0.87)
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.