Better Batches, Better Privacy: Random Allocation for DP-BandMF
cs.CR
Submitted: 2026-10-08
Updated: 2026-10-08
Code: https://github.com/oogle/grain
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Multi-Epoch Matrix Factorization Mechanisms for Private Machine Learning
- Privacy Amplification for BandMF via $b$-Min-Sep Subsampling
- Less Random, More Private: What is the Optimal Subsampling Scheme for DP-SGD?
- Efficient privacy loss accounting for subsampling and random allocation
- On Design Principles for Private Adaptive Optimizers
- Correlating Cross-Iteration Noise for DP-SGD using Model Curvature
- Exploring the Limits of Differentially Private Deep Learning with Group-wise Clipping
- DP- lambda CGD: Efficient Noise Correlation for Differentially Private Model Training
- Differentially Private Adaptive Optimization with Delayed Preconditioners
- AdaCliP: Adaptive Clipping for Private SGD
- Efficient and Scalable Implementation of Differentially Private Deep Learning without Shortcuts
- Sampling-Free Privacy Accounting for Matrix Mechanisms under Random Allocation
Related papers
- SoK: AI-Augmented Binary Reversing
- Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
- Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills
- Token Composition: A Graph Based on EVM Logs