Research map / Federated learning and privacy / Differential privacy
User-level Private Stochastic Convex Optimization with Optimal Rates
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 user-level DP (per-user averaged gradients with user-level DP mean estimation) in central model instead of Levy et al. (2021) approach | central DP user-level SCO · smooth convex loss, central DP | excess population risk: removes extra sqrt(d) factor relative to Levy et al. (2021) | Helped |
| Use accelerated mini-batch SGD with variance-reduced local DP gradient estimator instead of one-pass noisy SGD | local DP user-level SCO · smooth convex loss, local DP | excess population risk | Helped |
| Generalization properties of DP used to ensure concentration of adaptive local gradient estimates across users and iterations | central DP multi-pass SGD · smooth convex loss, central DP | excess population risk | Helped |
Closest papers by meaning
- Faster Algorithms for User-Level Private Stochastic Convex Optimization (NeurIPS 2024 · similarity 0.96)
- Differential Private Stochastic Optimization with Heavy-tailed Data: Towards Optimal Rates (AAAI 2025 · similarity 0.92)
- Private Stochastic Convex Optimization with Heavy Tails: Near-Optimality from Simple Reductions (NeurIPS 2024 · similarity 0.92)
- Public-data Assisted Private Stochastic Optimization: Power and Limitations (NeurIPS 2024 · similarity 0.91)
- Differentially Private Optimization with Sparse Gradients (NeurIPS 2024 · similarity 0.91)
- Improved Differentially Private and Lazy Online Convex Optimization: Lower Regret without Smoothness Requirements (ICML 2024 · similarity 0.90)
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.