Two new research papers explore optimized subsampling techniques for Differentially Private Stochastic Gradient Descent (DP-SGD). The first paper, focusing on random shuffling, provides tight upper and lower bounds within the f-DP framework, achieving near-ideal privacy with a high number of training rounds. The second paper introduces Balanced Iteration Subsampling (BIS), demonstrating that structured participation, rather than random sampling, leads to stronger privacy amplification and optimal trade-offs across noise spectrums. Evaluations show BIS consistently outperforms Poisson subsampling in low-noise regimes, reducing the required noise multiplier. AI
影响 These studies offer new methods for differentially private machine learning, potentially enabling higher utility models with stronger privacy guarantees.
排序理由 Two academic papers published on arXiv presenting novel theoretical and empirical results for DP-SGD subsampling techniques.
- arXiv
- Berry-Esseen theorem
- DP-SGD
- f-DP
- Balanced Iteration Subsampling (BIS)
- Murat Bilgehan Ertan
- Poisson subsampling
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
