Better Batches, Better Privacy: Random Allocation for DP-BandMF

본문 미리보기

arXiv:2610.12712v1 Announce Type: new Abstract: The current state of the art for batch selection and noise addition in differentially private training is a combination of DP-BandMF (informally, DP-SGD where the noise added in two different steps can be correlated if the steps are within $b$ of each other) and privacy amplification by sampling via Monte Carlo accounting. First, whereas prior batch selection strategies for DP-BandMF were inspired by Poisson sampling, random allocation is known to