Research map / Federated learning and privacy / Differential privacy
Sketch-Flip-Merge: Mergeable Sketches for Private Distinct Counting
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 randomized merging instead of deterministic xor merging (ablation) | SFM private PCSA sketches · BitcoinHeist, larger privacy budgets | relative efficiency: relative efficiency tending toward 4 | Helped |
| Use randomized merging instead of deterministic xor merging (ablation) | SFM private PCSA sketches · Private PCSA sketches | variance: up to a further 75% variance reduction | Helped |
| Use SFM (sym) rather than PS (loose) | Private distinct counting · BitcoinHeist | relative efficiency: 100× better efficiency | Helped |
| Use SFM (sym) rather than the PS estimators | Private distinct counting · BitcoinHeist | relative efficiency: at least an order of magnitude | Helped |
| Merge additional private summaries | SFM (sym) · Merged summaries at n=10^6 | RRMSE | Hurt |
| Use the composite likelihood estimator for cardinality | SFM summaries · Private distinct counting | RRMSE | Helped |
Closest papers by meaning
- Breaking the Quadratic Barrier: Robust Cardinality Sketches for Adaptive Queries (ICML 2025 · similarity 0.86)
- Unmasking Vulnerabilities: Cardinality Sketches under Adaptive Inputs (ICML 2024 · similarity 0.84)
- Distributed Differentially Private Data Analytics via Secure Sketching (ICML 2025 · similarity 0.83)
- Publishing Below-Threshold Triangle Counts under Local Weight Differential Privacy (NeurIPS 2026 · similarity 0.82)
- The Gaussian Mixing Mechanism: Renyi Differential Privacy via Gaussian Sketches (NeurIPS 2025 · similarity 0.82)
- Private Continual Counting of Unbounded Streams (NeurIPS 2025 · similarity 0.82)
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.