Paperlane

no. 14 · Friday 13 November 2026

Sharp Gaussian approximation for Polyak–Ruppert averaged linear stochastic approximation

Bogdan Butyrin, Eric Moulines, Alexey Naumov and Sergey Samsonov

math.PR · 26 pp

Abstract

We study how quickly the normalized Polyak–Ruppert average approaches its Gaussian limit in linear stochastic approximation with bounded i.i.d.\ coefficients. For step sizes decaying as $k^{-γ}$, with $1/2\leγ<1$, we prove an upper bound of order $n^{γ-1}$ on the approximation error over convex sets in fixed dimension. A matching lower bound holds for every fixed model. The fastest rate in this range is $n^{-1/2}$, attained at $γ=1/2$. We also obtain an $O(n^{-1/2})$ bound on the error of the recursive multiplier bootstrap procedure.

comments
26 pages
fields
Probability, Statistics Theory
pdf ↗ page arXiv:2611.15442 ↗