DP-λCGD: Efficient Noise Correlation for Differentially Private Model Training

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Kalinin, Nikita P., McKenna, Ryan, Pagh, Rasmus, Lampert, Christoph H.
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910209636564992
author Kalinin, Nikita P.
McKenna, Ryan
Pagh, Rasmus
Lampert, Christoph H.
author_facet Kalinin, Nikita P.
McKenna, Ryan
Pagh, Rasmus
Lampert, Christoph H.
contents Differentially private stochastic gradient descent (DP-SGD) is the gold standard for training machine learning models with formal differential privacy guarantees. Several recent extensions improve its accuracy by introducing correlated noise across training iterations. Matrix factorization mechanisms are a prominent example, but they correlate noise across many iterations and require storing previously added noise vectors, leading to substantial memory overhead in some settings. In this work, we propose a new noise correlation strategy that correlates noise only with the immediately preceding iteration and cancels a controlled portion of it. Our method relies on noise regeneration using a pseudorandom noise generator, eliminating the need to store past noise. As a result, it requires no additional memory beyond standard DP-SGD. We show that the computational overhead is minimal and empirically demonstrate improved accuracy over DP-SGD.
format Preprint
id arxiv_https___arxiv_org_abs_2601_22334
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DP-λCGD: Efficient Noise Correlation for Differentially Private Model Training
Kalinin, Nikita P.
McKenna, Ryan
Pagh, Rasmus
Lampert, Christoph H.
Machine Learning
Differentially private stochastic gradient descent (DP-SGD) is the gold standard for training machine learning models with formal differential privacy guarantees. Several recent extensions improve its accuracy by introducing correlated noise across training iterations. Matrix factorization mechanisms are a prominent example, but they correlate noise across many iterations and require storing previously added noise vectors, leading to substantial memory overhead in some settings. In this work, we propose a new noise correlation strategy that correlates noise only with the immediately preceding iteration and cancels a controlled portion of it. Our method relies on noise regeneration using a pseudorandom noise generator, eliminating the need to store past noise. As a result, it requires no additional memory beyond standard DP-SGD. We show that the computational overhead is minimal and empirically demonstrate improved accuracy over DP-SGD.
title DP-λCGD: Efficient Noise Correlation for Differentially Private Model Training
topic Machine Learning
url https://arxiv.org/abs/2601.22334