Research map / Federated learning and privacy / Communication-efficient FL
Breaking the Communication-Privacy-Accuracy Trilemma
Wei-Ning Chen, Peter Kairouz, Ayfer Ozgur
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 recursive Hadamard encoding to jointly meet privacy and communication constraints | Frequency estimation · Frequency estimation | ℓ2 error: RHR achieves \frac{d}{n\min\left\{ e^\varepsilon, \left( e^{\varepsilon} -1 \right)^2, 2^b, d\right\}} | Improved |
| Replace the full communication of HR with RHR | Frequency estimation · Truncated normalized Geo(0.8), d=10000, ε=5 | ℓ1 estimation error | Improved |
| Use RHR instead of HR | Frequency estimation · Uniform and truncated normalized Geo(0.8) | ℓ1 estimation error | No clear effect |
| Use Subset Selection instead of RHR | Frequency estimation · Uniform and truncated normalized Geo(0.8) | ℓ1 estimation error | Improved |
| Concatenate privatization and quantization rather than design them jointly | Mean estimation · Mean estimation | ℓ2 estimation error: d/n -> d^2 | Hurt |
| Concatenate Subset Selection and grouping-based quantization | Distribution estimation · Distribution estimation | ℓ2 error: O\left(\frac{d^2}{n2^b \min\left( e^\epsilon, (e^\epsilon-1)^2 \right)}\right) | Hurt |
Closest papers by meaning
- The Privacy Power of Correlated Noise in Decentralized Learning (ICML 2024 · similarity 0.86)
- Universal Exact Compression of Differentially Private Mechanisms (NeurIPS 2024 · similarity 0.86)
- Improved Communication-Privacy Trade-offs in $L_2$ Mean Estimation under Streaming Differential Privacy (ICML 2024 · similarity 0.86)
- Revisiting Gradient Pruning: A Dual Realization for Defending against Gradient Attacks (AAAI 2024 · similarity 0.85)
- A Huber Loss Minimization Approach to Mean Estimation under User-level Differential Privacy (NeurIPS 2024 · similarity 0.84)
- Learning from End User Data with Shuffled Differential Privacy over Kernel Densities (ICLR 2025 · similarity 0.84)
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.
