EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multi-iteration Stochastic Optimizers

André Carlon, Luis Espath, Rafael Holdorf, Raúl Tempone

Applied Mathematics & Optimization · 2026-06 · 93 (3)

DOI: 10.1007/s00245-026-10451-x ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Mathematics & Optimization, online erschienen am 29.05.2026. Verfasst von André Carlon und 3 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: King Abdullah University of Science and Technology, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We introduce Multi-Iteration Stochastic Optimizers, a novel class of first-order stochastic methods that control the relative $$L^2$$ L 2 error using successive control variates along the iteration path. By exploiting correlations between iterates, these control variates reduce the estimator’s variance, making an accurate mean gradient estimation computationally affordable. Our approach centers on the Multi-Iteration stochastiC Estimator (MICE), which can be seamlessly coupled with any first-order stochastic optimizer due to its non-intrusive design. The algorithm adaptively selects which iterates to include in its index set. We provide both an error analysis of MICE and a convergence analysis for Multi-Iteration Stochastic Optimizers across various problem classes, including some non-convex cases. In the smooth, strongly convex setting, we demonstrate that to approximate a minimizer within a tolerance tol , SGD-MICE requires, on average, $$O(tol^{-1})$$ O ( t o l - 1 ) stochastic gradient evaluations, compared to $$O(tol^{-1}\log (tol^{-1}))$$ O ( t o l - 1 log ( t o l - 1 ) ) for SGD with adaptive batch sizes. In numerical experiments, SGD-MICE achieved the desired tolerance with fewer than 3% of the gradient evaluations required by adaptive batch SGD. Additionally, MICE offers a straightforward stopping criterion based on the gradient norm, validated through consistency tests. To assess its efficiency, we present examples using both SGD-MICE and Adam-MICE, including a stochastic adaptation of the Rosenbrock function and logistic regression on various datasets. Compared to SGD, SAG, SAGA, SVRG, and SARAH, our approach consistently reduces the gradient sampling cost without the need for extensive parameter tuning.

Förderung

  • King Abdullah University of Science and Technology
  • Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

Zitieren

APA

Carlon, A., Espath, L., Holdorf, R., & Tempone, R. (2026). Multi-iteration Stochastic Optimizers. Applied Mathematics & Optimization, 93(3). https://doi.org/10.1007/s00245-026-10451-x

Vancouver

Carlon A, Espath L, Holdorf R, Tempone R. Multi-iteration Stochastic Optimizers. Applied Mathematics & Optimization. 2026;93(3). doi:10.1007/s00245-026-10451-x

BibTeX
@article{Carlon2026,
  author = {Carlon, André and Espath, Luis and Holdorf, Rafael and Tempone, Raúl},
  title = {Multi-iteration Stochastic Optimizers},
  journal = {Applied Mathematics & Optimization},
  year = {2026},
  volume = {93},
  number = {3},
  doi = {10.1007/s00245-026-10451-x}
}
Kostenlos starten →