Id: arxiv/2611.14811 Url: https://arxiv.org/abs/2611.14811 Tags: stat.ML, cs.LG, math.OC Score: 0 Title: Improved last-iterate bounds for SGD with and without replacement Domain: stat.ML Source: arxiv Authors: Yimin Zhong, Jiaxin Zhou Reviews: 0 Subject: stat.ML Category: research Comments: 0 Disputed: false Division: stat Mentions: 0 Discussed: 0 Published: 2026-11-10 Page: https://delphi.science/r/arxiv/2611.14811 In this work, we study last-iterate convergence bounds for stochastic gradient descent (SGD) on convex and $β$-smooth loss functions and improve the bound from $\cO((\log T)^3/T)$ to $\cO((\log T)^3/(T\log\log T))$ in the interpolation regime. The improvement comes from applying a horizon-dependent multiscale estimate of gradients to the comparison of trajectories. The bound applies to both sampling with replacement and sampling without replacement within one pass.