Research map / Federated learning and privacy / Differential privacy
Differentially Private Adaptive Optimization with Delayed Preconditioners
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 delayed preconditioners estimated by averaging prior private gradients, without auxiliary data | DP²-RMSProp · IMDB | accuracy: .687 -> .815 | Helped |
| Use delayed preconditioners estimated by averaging prior private gradients, without auxiliary data | DP²-RMSProp · StackOverflow | accuracy: .330 -> .391 | Helped |
| Use delayed preconditioners estimated by averaging prior private gradients, without auxiliary data | DP²-RMSProp · MovieLens | MSE: 3.02 -> 2.78 | Helped |
| Use moderate rather than very small or large delay | DP²-RMSProp · text and recommendation datasets | test performance | Mixed |
| Add delayed preconditioner updates while adding independent noise for both preconditioner accumulation and adaptive updates (ablation) | DP² on IMDB · IMDB | performance | Hurt |
| Apply the delayed preconditioner before privatizing the gradient, rather than after noising the clean gradient (ablation) | DP² on IMDB · IMDB | performance | Hurt |
Closest papers by meaning
- Differentially Private Domain Adaptation with Theoretical Guarantees (ICML 2024 · similarity 0.89)
- Differentially Private Optimization with Sparse Gradients (NeurIPS 2024 · similarity 0.89)
- Online Sensitivity Optimization in Differentially Private Learning (AAAI 2024 · similarity 0.88)
- Private Gradient Descent for Linear Regression: Tighter Error Bounds and Instance-Specific Uncertainty Estimation (ICML 2024 · similarity 0.88)
- Second-Order Convergence in Private Stochastic Non-Convex Optimization (NeurIPS 2025 · similarity 0.88)
- Differential Private Stochastic Optimization with Heavy-tailed Data: Towards Optimal Rates (AAAI 2025 · similarity 0.88)
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.