Adaptive Regularization within Trust Region Methods for Stochastic Nonconvex Optimization

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Ha, Yunsoo, Shashaani, Sara, Tran-dinh, Quoc
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866918473815293952
author Ha, Yunsoo
Shashaani, Sara
Tran-dinh, Quoc
author_facet Ha, Yunsoo
Shashaani, Sara
Tran-dinh, Quoc
contents We propose a stochastic nonconvex optimization algorithm that achieves almost sure $\tilde{\mathcal{O}}(ε^{-1.5})$ iteration complexity for problems with smooth objective functions and gradients only observable with noise. The mean-zero stochastic noise is decision-dependent and has unbounded support with subexponential tail, allowing our framework to cover a broad class of problems. The improved almost sure iteration complexity is achieved with a new variant of the adaptive sampling trust-region optimization (ASTRO) augmented with an adaptively regularized local model, which we term Reg-ASTRO. Adaptive sampling ensures that the estimation precision is aligned with a measure of stationarity, so that iterates closer to stationarity trigger higher accuracy requirement for sampling. A key analytical challenge arises because the trust-region radius and regularization are coupled and not determined prior to gradient estimation at each iteration. We further establish an almost sure $\tilde{\mathcal{O}}(ε^{-4.5})$ sample complexity for Reg-ASTRO, which improves to $\tilde{\mathcal{O}}(ε^{-3.5})$ under stronger regularity conditions and use of common random numbers, substantially outperforming first-order methods in theory and numerical experiments.
format Preprint
id arxiv_https___arxiv_org_abs_2604_15457
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Adaptive Regularization within Trust Region Methods for Stochastic Nonconvex Optimization
Ha, Yunsoo
Shashaani, Sara
Tran-dinh, Quoc
Optimization and Control
We propose a stochastic nonconvex optimization algorithm that achieves almost sure $\tilde{\mathcal{O}}(ε^{-1.5})$ iteration complexity for problems with smooth objective functions and gradients only observable with noise. The mean-zero stochastic noise is decision-dependent and has unbounded support with subexponential tail, allowing our framework to cover a broad class of problems. The improved almost sure iteration complexity is achieved with a new variant of the adaptive sampling trust-region optimization (ASTRO) augmented with an adaptively regularized local model, which we term Reg-ASTRO. Adaptive sampling ensures that the estimation precision is aligned with a measure of stationarity, so that iterates closer to stationarity trigger higher accuracy requirement for sampling. A key analytical challenge arises because the trust-region radius and regularization are coupled and not determined prior to gradient estimation at each iteration. We further establish an almost sure $\tilde{\mathcal{O}}(ε^{-4.5})$ sample complexity for Reg-ASTRO, which improves to $\tilde{\mathcal{O}}(ε^{-3.5})$ under stronger regularity conditions and use of common random numbers, substantially outperforming first-order methods in theory and numerical experiments.
title Adaptive Regularization within Trust Region Methods for Stochastic Nonconvex Optimization
topic Optimization and Control
url https://arxiv.org/abs/2604.15457