Research map / Federated learning and privacy / Differential privacy
Differentially Private Anonymized Histograms
Ananda Theertha Suresh
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 |
|---|---|---|---|
| Use geometric noise instead of Laplace noise | PrivHist low-privacy regime · Anonymized histograms, ε > 1 | expected ℓ1 noise: 2e^{-\epsilon}/(1-e^{-2\epsilon}) | Improved |
| Use prevalences rather than cumulative prevalences for small counts | PrivHist low-privacy regime · Anonymized histograms | sorted-ℓ1 error | Hurt |
| Split the histogram by threshold without sensitivity-preserving preprocessing | Histogram release at the small/large-count boundary · Anonymized histograms | privacy sensitivity: T + 2 | Hurt |
| Add geometric noise to the prevalences at T and T+1 before splitting | PrivHist · Anonymized histograms | privacy sensitivity | Improved |
| Use a single property-specific private estimator instead of releasing one private histogram for reuse | Symmetric-property estimation · Discrete distributions | sample complexity | Mixed |
Closest papers by meaning
- Anonymized Histograms in Intermediate Privacy Models (NeurIPS 2022 · similarity 0.95)
- Algorithms for bounding contribution for histogram estimation under user-level privacy (ICML 2023 · similarity 0.90)
- In Differential Privacy, There is Truth: on Vote-Histogram Leakage in Ensemble Private Learning (NeurIPS 2022 · similarity 0.88)
- INSPECTRE: Privately Estimating the Unseen (ICML 2018 · similarity 0.87)
- Learning discrete distributions: user vs item-level privacy (NeurIPS 2020 · similarity 0.87)
- A Generalized Shuffle Framework for Privacy Amplification: Strengthening Privacy Guarantees and Enhancing Utility (AAAI 2024 · similarity 0.85)
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.
