arXiv:2610. 01599v1 Announce Type: cross Abstract: 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.
arXiv:2610. 13799v1 Announce Type: cross Abstract: Variance reduction improves the convergence rate of stochastic non-convex optimization under the average smoothness condition.
arXiv:2609. 12785v1 Announce Type: new Abstract: Classical convergence guarantees for stochastic gradient methods typically assume Lipschitz-smooth objectives and finite-variance gradient noise, both frequently violated in practice.
The paper investigates variance reduction techniques for non‑convex optimization when gradient evaluations are delayed. It introduces an ordered recursive momentum estimator whose oracle complexity depends on smoothness, variance, and delay, and provides bounds for both stochastic and finite‑sum settings. An asynchronous snapshot variant is also proposed, allowing updates during full‑gradient computations under a completion‑time bound.
The paper introduces RanSOM, a second‑order momentum algorithm that eliminates the need for a Lipschitz Hessian or auxiliary forward/backward passes by randomizing the step length. By sampling the step from an exponential distribution and using a Stein‑type identity, the method obtains an unbiased estimate of the gradient change with a single Hessian‑vector product that also serves the next gradient computation. RanSOM-E achieves optimal oracle‑complexity exponents for heavy‑tailed noise without clipping, while RanSOM-B preserves feasibility for Frank‑Wolfe iterates, and experiments on CIFAR‑10 show comparable accuracy to existing methods with lower update time.
arXiv:2610. 33381v2 Announce Type: new Abstract: Training is often stopped at a data-dependent time, so guarantees for the last iterate must hold at all times simultaneously.
arXiv:2610. 13789v1 Announce Type: new Abstract: Momentum methods such as STORM and second-order momentum track the gradient by adding a transport correction to momentum, at the cost of an extra oracle call per iteration.
arXiv:2609. 29458v1 Announce Type: new Abstract: We present a tightened convergence analysis of clipped gradient descent on $(L_0, L_1)$-smooth functions, with quantitative constants.
arXiv:2504. 09951v2 Announce Type: replace-cross Abstract: We revisit a classical assumption for analyzing stochastic gradient algorithms where the squared norm of the stochastic subgradient (or the variance for smooth problems) is allowed to grow as fast as the squared norm of the optimization variable.
arXiv:2610.07814v1 Announce Type: cross Abstract: How far can stochastic gradient descent ascent (SGDA) go by tuning its timescale ratio and step sizes in nonconvex min-max games? We answer this ques...
arXiv:2505.20817v3 Announce Type: replace-cross Abstract: Gradient clipping is widely used in language-model training to control heavy-tailed gradient noise and can improve convergence guarantees ove...
The paper establishes the first optimal last‑iterate convergence rates, in high probability, for Clipped‑SGD when the gradient noise is heavy‑tailed. It addresses the gap left by earlier results that required strong noise assumptions, such as exponentially decaying tails. By proving these rates, the authors provide theoretical support for the practical effectiveness of gradient clipping in stochastic convex optimization.
arXiv:2508.04860v2 Announce Type: replace-cross Abstract: Stochastic Gradient Descent (SGD) is a cornerstone of large-scale optimization, yet its theoretical behavior under heavy-tailed noise -- comm...