Self-normalized Cramér-type moderate deviation of stochastic gradient Langevin dynamics
Journal of Applied Probability · 2026-09 · 63 (3) · 1034-1061
DOI: 10.1017/jpr.2025.10065 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Applied Probability, online erschienen am 02.03.2026. Verfasst von Hongsheng Dai und 2 weiteren Autor:innen. Die Arbeit nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
In this paper, we study the self-normalized Cramér-type moderate deviation of the empirical measure of the stochastic gradient Langevin dynamics (SGLD). Consequently, we also derive the Berry–Esseen bound for the SGLD. Our approach is by constructing a stochastic differential equation to approximate the SGLD and then applying Stein’s method to decompose the empirical measure into a martingale difference series sum and a negligible remainder term.
Zitieren
APA
Dai, H., Fan, X., & Lu, J. (2026). Self-normalized Cramér-type moderate deviation of stochastic gradient Langevin dynamics. Journal of Applied Probability, 63(3), 1034-1061. https://doi.org/10.1017/jpr.2025.10065
Vancouver
Dai H, Fan X, Lu J. Self-normalized Cramér-type moderate deviation of stochastic gradient Langevin dynamics. Journal of Applied Probability. 2026;63(3):1034-1061. doi:10.1017/jpr.2025.10065
BibTeX
@article{Dai2026,
author = {Dai, Hongsheng and Fan, Xiequan and Lu, Jianya},
title = {Self-normalized Cramér-type moderate deviation of stochastic gradient Langevin dynamics},
journal = {Journal of Applied Probability},
year = {2026},
volume = {63},
number = {3},
pages = {1034--1061},
doi = {10.1017/jpr.2025.10065}
}
EUVIMED