On the Rate of Gaussian Approximation for Linear Regression Problems

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Khusainov, Marat, Sheshukova, Marina, Durmus, Alain, Samsonov, Sergey
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916954818740224
author Khusainov, Marat
Sheshukova, Marina
Durmus, Alain
Samsonov, Sergey
author_facet Khusainov, Marat
Sheshukova, Marina
Durmus, Alain
Samsonov, Sergey
contents In this paper, we consider the problem of Gaussian approximation for the online linear regression task. We derive the corresponding rates for the setting of a constant learning rate and study the explicit dependence of the convergence rate upon the problem dimension $d$ and quantities related to the design matrix. When the number of iterations $n$ is known in advance, our results yield the rate of normal approximation of order $\sqrt{\log{n}/n}$, provided that the sample size $n$ is large enough.
format Preprint
id arxiv_https___arxiv_org_abs_2509_14039
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle On the Rate of Gaussian Approximation for Linear Regression Problems
Khusainov, Marat
Sheshukova, Marina
Durmus, Alain
Samsonov, Sergey
Machine Learning
Optimization and Control
60F05, 62L20, 93E35
In this paper, we consider the problem of Gaussian approximation for the online linear regression task. We derive the corresponding rates for the setting of a constant learning rate and study the explicit dependence of the convergence rate upon the problem dimension $d$ and quantities related to the design matrix. When the number of iterations $n$ is known in advance, our results yield the rate of normal approximation of order $\sqrt{\log{n}/n}$, provided that the sample size $n$ is large enough.
title On the Rate of Gaussian Approximation for Linear Regression Problems
topic Machine Learning
Optimization and Control
60F05, 62L20, 93E35
url https://arxiv.org/abs/2509.14039