Research map / Federated learning and privacy / Differential privacy
Lossless Compression of Efficient Private Local Randomizers
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 |
|---|---|---|---|
| Replace full RAPPOR bit-vector reports with pairwise-independent affine-function reports | RAPPOR for frequency estimation · Frequency estimation | variance per count | No clear effect |
| Replace full RAPPOR bit-vector reports with pairwise-independent affine-function reports | RAPPOR for frequency estimation · Frequency estimation | communication: k bits -> 2 \lceil \log_2 p\rceil bits | Helped |
| Compress an efficient LDP randomizer using PRG-generated randomness and rejection sampling | Efficient LDP randomizers · General LDP protocols | communication: linear in the dimension -> O(\log(T+T(\mathcal{R},G,\gamma)) | Helped |
| Compress PrivUnit using the general PRG-based method | PrivUnit · Mean estimation | communication: \Omega(d) -> O(\log d + \eps) bits | Helped |
| Use the simple low-communication randomizer after reducing the ε > 1 case to ε < 1 | Mean estimation · Mean estimation | running time: O(e^\eps d) -> O( \lceil \eps \rceil d) | Helped |
| Use Kashin-representation-based low-communication mean estimation instead of PrivUnit | Mean estimation · Mean estimation | variance: nearly a factor of 5\times larger | Hurt |
Closest papers by meaning
- Universal Exact Compression of Differentially Private Mechanisms (NeurIPS 2024 · similarity 0.89)
- Locally Optimal Private Sampling: Beyond the Global Minimax (NeurIPS 2025 · similarity 0.87)
- Improved Communication-Privacy Trade-offs in $L_2$ Mean Estimation under Streaming Differential Privacy (ICML 2024 · similarity 0.86)
- The Privacy Power of Correlated Noise in Decentralized Learning (ICML 2024 · similarity 0.86)
- PREAMBLE: Private and Efficient Aggregation via Block Sparse Vectors (NeurIPS 2025 · similarity 0.86)
- Lightweight Protocols for Distributed Private Quantile Estimation (ICML 2025 · 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.