Metraipolis › archive › dc_f39a65b7caae
Large-Step Last-Iterate Convergence of Random-Reshuffling Stochastic Gradient Descent — arXiv:2610.19444
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.19444
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T14:23:22.499Z
- SHA-256 of the archived text
c2ac21bab6766c4d7430468af3c311196d90c7c20145c089e6239e30ee411c9d- Stable URL of this page
https://metraipolis.com/documento/dc_f39a65b7caae
How to cite
Large-Step Last-Iterate Convergence of Random-Reshuffling Stochastic Gradient Descent — arXiv:2610.19444 — arXiv (en), https://arxiv.org/abs/2610.19444 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f39a65b7caae (Metraipolis archive dc_f39a65b7caae, sha256 c2ac21bab6766c4d7430468af3c311196d90c7c20145c089e6239e30ee411c9d)
BibTeX
@misc{dcf39a65b7caae,
title = {Large-Step Last-Iterate Convergence of Random-Reshuffling Stochastic Gradient Descent — arXiv:2610.19444},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.19444},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f39a65b7caae; sha256 c2ac21bab6766c4d7430468af3c311196d90c7c20145c089e6239e30ee411c9d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19444. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Random reshuffling stochastic gradient descent (RR-SGD) is widely used for finite-sum optimization in machine learning. For convex $L$-smooth components, recent work establishes convergence guarantees for an averaged iterate with component stepsizes as large as order $1/L$. However, the final iterate is a natural output in practice, and existing last-iterate guarantees in this setting require stepsizes that scale inversely with the number of components $n$. We establish an expected objective-gap bound for the last iterate of standard RR-SGD on unconstrained problems with a constant component stepsize satisfying $0<\eta\lesssim 1/(L(1 + \log(nK)))$, where $K$ is the number of epochs. By retaining the sampling bias at random comparison points, we remove the explicit $1/n$ factor from the stepsize restriction. Combined with existing small-step bounds, our result matches the corresponding averaged-iterate convergence rate up to logarithmic factors. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Optimization and Control (math.OC) ; Machine Learning (cs.LG); Machine Learning (stat.ML). Pagina del lavoro: https://arxiv.org/abs/2610.19444 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T13:53:35.858Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).