AI RESEARCH
Beyond Bounded Variance: Variance-Reduced Normalized Methods for Nonconvex Optimization under Blum-Gladyshev Noise
arXiv CS.LG
•
ArXi:2605.15314v1 Announce Type: new We study nonconvex stochastic optimization under the Blum-Gladyshe ($\mathsf{BG}$-0) noise model, where the stochastic gradient variance grows quadratically with the distance from the initialization. We consider this problem under both standard smoothness and the symmetric generalized-smoothness framework, which captures objectives whose local curvature can scale with the gradient norm.