English(EN) Settling the Burn-in Gap: Sharp Risk Bounds for Tail-Averaged SGD in Linear Regression

线性回归中尾部平均 SGD 的新风险界限

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-11 05:00

本文介绍了一个新的理论框架，用于分析高维线性回归中带有尾部平均的随机梯度下降 (SGD) 的性能。作者们在超额风险上建立了匹配的上界和下界，通过考虑初始化对谱性质的影响来完善现有分析。他们证明了初始误差的影响在预热阶段呈指数衰减，并确定了满足其理论条件的特定步长范围。 AI

影响完善了优化算法的理论理解，可能导致更有效的机器学习模型训练。

排序理由该条目是一篇学术论文，详细介绍了机器学习算法的理论进展。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 34 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该条目是一篇学术论文，详细介绍了机器学习算法的理论进展。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, other; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

arXiv cs.LG TIER_1 English(EN) · Qingyue Zhao, Xuheng Li, Quanquan Gu · 2026-11-11 05:00

弥合烧录差距：线性回归中尾部平均 SGD 的尖锐风险界限

arXiv:2611.13383v1 Announce Type: new Abstract: We study the excess risk of constant-stepsize online stochastic gradient descent (SGD) with tail-averaging for high-dimensional linear regression. Building on the instance-dependent spectral framework of Zou et al. (2021), we establ…