Research map / Federated learning and privacy / Differential privacy
Differentially Private Estimation and Inference in High-Dimensional Regression with FDR Control
Zhanrui Cai, Sai Li, Xintao Xia, Linjun Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add finite-sample correction to DP debiased CI (ablation) | DP debiased procedure · Toeplitz design, rho=0.0 | average coverage: 0.823 -> 0.951 | Improved |
| add finite-sample correction to DP debiased CI (ablation) | DP debiased procedure · Toeplitz design, rho=0.6 | average coverage: 0.896 -> 0.964 | Improved |
| add finite-sample correction to DP debiased CI (ablation) | DP debiased procedure · Block equicorrelated design, rho=0.20 | average coverage: 0.820 -> 0.945 | Improved |
| add finite-sample correction to DP debiased CI (ablation) | DP debiased procedure · Toeplitz design, rho=0.2 | average CI length: 0.089 -> 0.127 | Hurt |
| add finite-sample correction to DP debiased CI (ablation) | DP debiased procedure · Block equicorrelated design, rho=0.10 | average CI length: 0.088 -> 0.127 | Hurt |
| apply DP debiased inference instead of non-private debiased Lasso | DP correction · Toeplitz design, rho=0.2 | average CI length: 0.089 -> 0.127 | Hurt |
| increase sample size n for DP-FDR control | DP-FDR control · FDR control, p=10000, rho=0.2, xi=0.3 | empirical FDR and power | Improved |
| use DP-FDR control instead of non-private FDR control | mirror-statistics FDR control · FDR control, n=p=10000, epsilon=4 | power | Hurt |
Closest papers by meaning
- Better Locally Private Sparse Estimation Given Multiple Samples Per User (ICML 2024 · similarity 0.89)
- Improved Analysis of Sparse Linear Regression in Local Differential Privacy Model (ICLR 2024 · similarity 0.89)
- On the Computational Complexity of Private High-dimensional Model Selection (NeurIPS 2024 · similarity 0.89)
- Differentially Private High-dimensional Variable Selection via Integer Programming (NeurIPS 2025 · similarity 0.88)
- On Inherent Privacy of Posterior Sampling: A Unified R\'enyi-Divergence Framework (NeurIPS 2026 · similarity 0.88)
- Truthful High Dimensional Sparse Linear Regression (NeurIPS 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.
