Showing math.OCShow all
3 papers · 1 filter
math.OC2026
The Last Iterate of the Projected Stochastic Subgradient Method Pays a Logarithm, Even in Dimension One
Guglielmo Beretta, Tommaso Cesari, Roberto Colomboni
We prove that the projected stochastic subgradient method with the standard anytime stepsizes can suffer expected last-iterate optimization error of order $(\log…
math.OC2026
A Minimax Analysis of the Last Iterate of the SubGradient Method in Finite Dimension
Guglielmo Beretta, Tommaso Cesari, Roberto Colomboni +1
We study the last iterate of the projected subGradient Method with a constant stepsize for convex -Lipschitz objectives in , minimized over closed convex sets. Ass…
math.OC2026
New Bounds for the Last Iterate of the Stochastic subGradient Method
Guglielmo Beretta, Tommaso Cesari, Roberto Colomboni +1
We study the last iterate of the stochastic subgradient method for one-dimensional convex Lipschitz objectives. For a fixed horizon , we consider the standard fixed stepsizes $η…