{"id":"arxiv/2610.01599","url":"https://arxiv.org/abs/2610.01599","tags":["math.OC","cs.LG"],"score":0,"title":"Convergence Analysis of STORM Under Different Geometries","domain":"math.OC","source":"arxiv","authors":["Wei Jiang","Yibo Wang","Wenhao Yang","Rui Yan","Lijun Zhang","Zechao Li"],"reviews":0,"subject":"math.OC","summary":"Stochastic recursive momentum (STORM) achieves fast convergence for nonconvex optimization via the variance reduction effect, but existing analyses rely on the strong average smoothness assumption. In this paper, we study the convergence of STORM for different objectives without average smoothness. We first revisit the results under average smoothness, obtaining the $O(T^{-1/3})$ bound for nonconvex objectives and the $O(σ^2/(μT))$ bound for last-iterate output under the $μ$-Polyak--Łojasiewicz~(PL) condition. Without average smoothness, we design an auxiliary sequence and compare the STORM…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-01","reactions":{}}