{"id":"arxiv/2611.02534","url":"https://arxiv.org/abs/2611.02534","tags":["stat.CO","math.OC"],"score":0,"title":"Subgradient Methods for Nonsmooth Sample Average Approximation","domain":"stat.CO","source":"arxiv","authors":["Xi Chen","Ouqi Fu","Jiyuan Tu","Yichen Zhang"],"reviews":0,"subject":"stat.CO","summary":"We study direct subgradient methods for nonsmooth sample average approximation (SAA). Subgradient descent and Nesterov acceleration with constant step sizes attain polynomial and geometric rates matching their smooth counterparts. After $O(\\log n)$ full-batch updates, where $n$ is the sample size, the computed solution has the same first-order expansion and limiting distribution as an SAA solution, even when the iterates do not converge. A bound on subgradient increments transfers population regularity to the sample objective, controls the algorithmic trajectory, and identifies the leading…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-11-02","reactions":{}}