{"id":"arxiv/2610.42071","url":"https://arxiv.org/abs/2610.42071","tags":["math.OC","math.NA","math.PR"],"score":0,"title":"Stochastic Rayleigh-Ritz Ascent: Global Linear Convergence under a Spectral Gap","domain":"math.OC","source":"arxiv","authors":["Jonas Bresch","Oleh Melnyk"],"reviews":0,"subject":"math.OC","summary":"We consider the adjoint-free random-plane iteration algorithm for finding the top eigenvector of a matrix proposed in (SIAM J. Matrix Anal. Appl. 47 (2026), 1349-1374). Unlike standard methods such as the power method or gradient descent, the above algorithm lacked linear convergence guarantees in the presence of a spectral gap. In this paper, we establish explicit spectral-gap-dependent global linear convergence rates in expectation, providing quantitative guarantees for the empirically observed convergence.","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-30","reactions":{}}