Research map / Federated learning and privacy / Differential privacy
Muffliato: Peer-to-Peer Privacy Amplification for Decentralized Optimization and Averaging
Edwige Cyffers, Mathieu Even, Aurélien Bellet, Laurent Massoulié
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 |
|---|---|---|---|
| add gossip averaging after local noise injection (Muffliato) versus LDP | Muffliato · synthetic exponential graph, n=2048 | pairwise privacy loss | Improved |
| use Erdos-Renyi graph versus exponential graph | Muffliato · synthetic graphs, n=2048 | privacy-utility trade-off | No clear effect |
| use grid topology versus exponential or Erdos-Renyi | Muffliato · synthetic graphs, n=2048 | privacy-utility trade-off | Hurt |
| increase graph distance between nodes | Muffliato · Facebook ego graph | pairwise privacy loss | Improved |
| use time-varying Erdos-Renyi graphs after each gradient step | Muffliato-GD · UCI Housing logistic regression | privacy loss uniformity across nodes | Improved |
| increase number of nodes from n=2000 to n=4000 | Muffliato-GD · UCI Housing logistic regression | privacy gain over LDP | Improved |
| replace trusted central server aggregation with gossiping | Muffliato-GD · UCI Housing logistic regression | accuracy | No clear effect |
Closest papers by meaning
- Differentially Private Decentralized Learning with Random Walks (ICML 2024 · similarity 0.92)
- The Privacy Power of Correlated Noise in Decentralized Learning (ICML 2024 · similarity 0.91)
- Mitigating the Privacy–Utility Trade-off in Decentralized Federated Learning via f-Differential Privacy (NeurIPS 2025 · similarity 0.89)
- Revisiting Differentially Private Algorithms for Decentralized Online Learning (ICML 2025 · similarity 0.89)
- Privacy amplification by random allocation (NeurIPS 2025 · similarity 0.88)
- Locally Differentially Private Decentralized Stochastic Bilevel Optimization with Guaranteed Convergence Accuracy (ICML 2024 · similarity 0.87)
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.
