Research map / Federated learning and privacy / Differential privacy
Differentially Private Chi-Squared Hypothesis Testing: Goodness of Fit and Independence Testing
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 |
|---|---|---|---|
| Account for privacy-noise distribution when setting the goodness-of-fit rejection threshold | GOF on private counts · GOF d=100, n=1,500 | significance: 0.0000 -> $0.9522$ | Helped |
| Account for privacy-noise distribution when setting the goodness-of-fit rejection threshold | GOF on private counts · GOF d=100, n=10,000 | significance: 0.0000 -> $0.9491$ | Helped |
| Account for privacy-noise distribution when setting the goodness-of-fit rejection threshold | GOF on private counts · GOF d=100, n=100,000 | significance: 0.0000 -> $0.9511$ | Helped |
| Account for privacy-noise distribution when setting the goodness-of-fit rejection threshold | GOF on private counts · GOF d=100, n=1,000,000 | significance: 0.0524 -> $0.9479$ | Helped |
| Use privacy-calibrated tests instead of classical tests on private counts | Goodness-of-fit and independence testing · Simulated data under the null hypothesis | significance | Helped |
| Add Laplace or Gaussian privacy noise to counts | Classical non-private tests · Goodness-of-fit and independence testing | power | Hurt |
| Use Laplace noise rather than Gaussian noise at the same privacy level | Differentially private chi-squared tests · Goodness-of-fit and independence testing | power | Helped |
| Increase sample size when using Laplace noise | Differentially private independence tests · 2×2 contingency tables | power: less than 3000 additional samples | Helped |
Closest papers by meaning
- Local Private Hypothesis Testing: Chi-Square Tests (ICML 2018 · similarity 0.86)
- Priv’IT: Private and Sample Efficient Identity Testing (ICML 2017 · similarity 0.86)
- Optimal Private and Communication Constraint Distributed Goodness-of-Fit Testing for Discrete Distributions in the Large Sample Regime (NeurIPS 2024 · similarity 0.80)
- Minimax Optimal Two-Sample Testing under Local Differential Privacy (NeurIPS 2026 · similarity 0.79)
- Nearly-Linear Time Private Hypothesis Selection with the Optimal Approximation Factor (NeurIPS 2025 · similarity 0.79)
- Optimal Rates for Differentially Private Hypothesis Testing with E-values (NeurIPS 2026 · similarity 0.78)
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.