Research map / Federated learning and privacy / Differential privacy
DP-Fast MH: Private, Fast, and Accurate Metropolis-Hastings for Large-Scale Bayesian Inference
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 DP-Fast MH instead of the penalty algorithm | Private MH for truncated Gaussian mixture · Truncated Gaussian Mixture | symmetric KL | Helped |
| use DP-Fast MH instead of the Barker method | Private MH for truncated Gaussian mixture · Truncated Gaussian Mixture | symmetric KL: 4000 iterations; 15 seconds; 10^4 iterations; 20 seconds | Helped |
| use DP-Fast MH instead of previous private methods | Private logistic regression · MNIST 7s vs. 9s | test accuracy | Helped |
| use DP-Fast MH minibatches instead of the full-batch version (ablation) | DP-Fast MH · Truncated Gaussian Mixture | symmetric KL | No clear effect |
| use DP-Fast MH minibatches instead of the full-batch version (ablation) | DP-Fast MH · MNIST 7s vs. 9s | test accuracy | No clear effect |
| decrease the minibatch upper bound K (ablation) | DP-Fast MH logistic regression · MNIST 7s vs. 9s | test accuracy | Helped |
| decrease the minibatch upper bound K (ablation) | DP-Fast MH logistic regression · MNIST 7s vs. 9s | batch size | Helped |
| use the full-batch version instead of the minibatch version (ablation) | DP-Fast MH · Truncated Gaussian Mixture | convergence time | Hurt |
Closest papers by meaning
- Tractable MCMC for Private Learning with Pure and Gaussian Differential Privacy (ICLR 2024 · similarity 0.91)
- Differential Privacy Guarantees of Markov Chain Monte Carlo Algorithms (ICML 2025 · similarity 0.90)
- On Inherent Privacy of Posterior Sampling: A Unified R\'enyi-Divergence Framework (NeurIPS 2026 · similarity 0.90)
- Privacy Guarantees in Posterior Sampling under Contamination (NeurIPS 2026 · similarity 0.88)
- On the Computational Complexity of Private High-dimensional Model Selection (NeurIPS 2024 · similarity 0.88)
- Locally Optimal Private Sampling: Beyond the Global Minimax (NeurIPS 2025 · 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.