Research map / Federated learning and privacy / Differential privacy
New Oracle-Efficient Algorithms for Private Synthetic Data Release
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 rejection sampling to recycle samples from previous rounds (ablation) | DQ · Private synthetic data release | accuracy error bound: $$\alpha = O\left(\frac{\log(|\mathcal{X}|T/\beta) \cdot \log^{3}(|\mathcal{Q}|) \cdot \log(1/\delta)}{n^{2} \varepsilon | Helped |
| Use Separator-FTPL instead of Non-Convex-FTPL, assuming a small separator set (ablation) | FEM · Private synthetic data release with a small-separator query class | accuracy error bound: $\widetilde O\left( d^{5/8} \log^{1/2}|\mathcal{Q}| / n^{1/2}\right)$ versus $\widetilde O\left(d^{3/4} \log^{1/2}|\math | Helped |
| Use FEM rather than HDMM | Private synthetic data release · ADULT and LOANS, 3-way and 5-way marginals | max additive error | Helped |
Closest papers by meaning
- Iterative Methods for Private Synthetic Data: Unifying Framework and New Methods (NeurIPS 2021 · similarity 0.92)
- Generating Private Synthetic Data with Genetic Algorithms (ICML 2023 · similarity 0.90)
- Oracle-Efficient Differentially Private Learning with Public Data (NeurIPS 2024 · similarity 0.89)
- Privacy-Preserving Data Release Leveraging Optimal Transport and Particle Gradient Descent (ICML 2024 · similarity 0.88)
- Private Evolution Converges (NeurIPS 2025 · similarity 0.88)
- Privacy Amplification Persists under Unlimited Synthetic Data Release (NeurIPS 2026 · similarity 0.88)
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.