ExtraSAGA: Variance Reduction Hybrid Method for Variational Inequalities
Doklady Mathematics · 2025-12 · 112 (3) · 527-540
DOI: 10.1134/s1064562425700607 ↗Kurz gefasst
Fachartikel in der Zeitschrift Doklady Mathematics, online erschienen am 22.03.2026. Verfasst von G. Chirkov und 5 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Variational inequalities (VIs) serve as a powerful tool for various problems. This setting can be applied to a wide range of optimization, machine learning (ML), and other challenges. At the same time, large volumes of data are essential for high performance in ML tasks, which are addressed through stochastic approaches. However, widely used SGD method suffers from a non-decreasing variance of the stochastic gradient. To resolve this issue, variance reduction techniques were developed. This approach is well-studied for minimization but less extensively for VIs. In this paper, we modify the SAGA method, known for its effectiveness in stochastic minimization, by integrating Extragradient to address VI problems. We provide a theoretical analysis of the proposed method and conduct experiments, including bilinear problems and image denoising tasks.
Zitieren
APA
Chirkov, G., Kabikov, Y., Medyakov, D., Molodtsov, G., Shestakov, A., & Beznosikov, A. (2025). ExtraSAGA: Variance Reduction Hybrid Method for Variational Inequalities. Doklady Mathematics, 112(3), 527-540. https://doi.org/10.1134/s1064562425700607
Vancouver
Chirkov G, Kabikov Y, Medyakov D, Molodtsov G, Shestakov A, Beznosikov A. ExtraSAGA: Variance Reduction Hybrid Method for Variational Inequalities. Doklady Mathematics. 2025;112(3):527-540. doi:10.1134/s1064562425700607
BibTeX
@article{Chirkov2025,
author = {Chirkov, G. and Kabikov, Yu. and Medyakov, D. and Molodtsov, G. and Shestakov, A. and Beznosikov, A.},
title = {ExtraSAGA: Variance Reduction Hybrid Method for Variational Inequalities},
journal = {Doklady Mathematics},
year = {2025},
volume = {112},
number = {3},
pages = {527--540},
doi = {10.1134/s1064562425700607}
}
EUVIMED