Convergence of Stochastic Gradient Methods under Heavy-Tailed Noise and Hölder Smoothness

• URL: http://arxiv.org/abs/2609.12785v1
• Date: Fri, 11 Sep 2026 12:40:43 GMT
• Title: Convergence of Stochastic Gradient Methods under Heavy-Tailed Noise and Hölder Smoothness
• Abstract summary: We analyze gradient clipping (G-Clip) and show that it recovers the rate above for $ge1+s$ while in the very heavy-tailed regime regime it has $1+s$.
• Score: 0.0
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Classical convergence guarantees for stochastic gradient methods typically assume Lipschitz-smooth objectives and finite-variance gradient noise, both frequently violated in practice. In contrast, we study nonconvex stochastic optimization under the joint relaxation of these assumptions: objectives with $(L,s)$-Hölder continuous gradients, $s\in(0,1]$, and gradient noise satisfying only a bounded $α$-th moment condition for $α\in(1,2]$. We establish three convergence results. Firstly, that standard SGD converges at rate $O(T^{-s/(1+s)})$ whenever $α\ge1+s$, extending the classical nonconvex SGD rate to heavy-tailed noise and Hölder smoothness simultaneously. Secondly, we analyze $δ$-regularized gradient clipping ($δ$-GClip), a provable trainer of wide and deep nets, and establish a stationarity rate of $O(T^{-2s(α-1)/[(1+s)(2α-1)]})$ under the same condition. Thirdly, we analyze standard gradient clipping (G-Clip) and show that it recovers the above rate for $α\ge1+s$ while in the very heavy-tailed regime $α<1+s$, it has a convergence rate $O(T^{-2s(α-1)/[(α-1)+s(2α-1)]})$ --- the first convergence guarantee in this regime for any stochastic gradient based method.

Related papers

• Beyond Bounded Variance: Variance-Reduced Normalized Methods for Nonconvex Optimization under Blum-Gladyshev Noise [7.692336118507715]
  We study non-smoother optimization under the Blum-Gladyshev ($mathsfBG$-0) noise model.<n>We prove that normalized gradient descent with momentum, using only one parameter per gradient, converges under $mathsfBG$-0 noise with complexity $O(varepsilon-6)$.
  arXiv  Detail & Related papers  (2026-05-14T18:27:49Z)
• When Does $\ell_2$-Boosting Overfit Benignly? High-Dimensional Risk Asymptotics and the $\ell_1$ Implicit Bias [15.113649527486276]
  We show that benign overfitting fails at the linear rate.<n>We show that while this localization mechanism should persist in the presence of signals, the exact signal-noise decomposition remains an open problem.
  arXiv  Detail & Related papers  (2026-05-07T14:14:09Z)
• Complexity Bounds for Smooth Convex Multiobjective Optimization [0.0]
  We study the oracle complexity of finding $varepsilon$-Pareto stationary points in smooth multiobjective optimization with $m$ objectives.<n>For strongly convex objectives, any span first-order method exhibits linear convergence no faster than $exp(-Theta(T/sqrtkappa)$ after $T$ oracle calls.<n>For convex problems and general span methods with adaptive scalarizations, we establish a universal lower bound of order $1/T2$ on the gradient norm of the final iterate after $T$ steps.
  arXiv  Detail & Related papers  (2025-09-16T21:33:11Z)
• Generalized Gradient Norm Clipping & Non-Euclidean $(L_0,L_1)$-Smoothness [51.302674884611335]
  This work introduces a hybrid non-Euclidean optimization method which generalizes norm clipping by combining steepest descent and conditional gradient approaches.<n>We discuss how to instantiate the algorithms for deep learning and demonstrate their properties on image classification and language modeling.
  arXiv  Detail & Related papers  (2025-06-02T17:34:29Z)
• Nonconvex Stochastic Optimization under Heavy-Tailed Noises: Optimal Convergence without Gradient Clipping [21.865728815935665]
  We provide the first convergence under heavy-tailed noises but without clipping.<n>We also establish first $mathcalO(Tfrac1-mathfrakp3mathfrakp-2)$ convergence rate in the case where the tail index $mathfrakp$ is unknown in advance.
  arXiv  Detail & Related papers  (2024-12-27T08:46:46Z)
• On the $O(\frac{\sqrt{d}}{T^{1/4}})$ Convergence Rate of RMSProp and Its Momentum Extension Measured by $\ell_1$ Norm [54.28350823319057]
  This paper considers the RMSProp and its momentum extension and establishes the convergence rate of $frac1Tsum_k=1T.<n>Our convergence rate matches the lower bound with respect to all the coefficients except the dimension $d$.<n>Our convergence rate can be considered to be analogous to the $frac1Tsum_k=1T.
  arXiv  Detail & Related papers  (2024-02-01T07:21:32Z)
• Optimal Extragradient-Based Bilinearly-Coupled Saddle-Point Optimization [116.89941263390769]
  We consider the smooth convex-concave bilinearly-coupled saddle-point problem, $min_mathbfxmax_mathbfyF(mathbfx) + H(mathbfx,mathbfy)$, where one has access to first-order oracles for $F$, $G$ as well as the bilinear coupling function $H$. We present a emphaccelerated gradient-extragradient (AG-EG) descent-ascent algorithm that combines extragrad
  arXiv  Detail & Related papers  (2022-06-17T06:10:20Z)
• Sharper Convergence Guarantees for Asynchronous SGD for Distributed and Federated Learning [77.22019100456595]
  We show a training algorithm for distributed computation workers with varying communication frequency. In this work, we obtain a tighter convergence rate of $mathcalO!!!(sigma2-2_avg!! . We also show that the heterogeneity term in rate is affected by the average delay within each worker.
  arXiv  Detail & Related papers  (2022-06-16T17:10:57Z)
• Perseus: A Simple and Optimal High-Order Method for Variational Inequalities [81.32967242727152]
  A VI involves finding $xstar in mathcalX$ such that $langle F(x), x - xstarrangle geq 0$ for all $x in mathcalX$. We propose a $pth$-order method that does textitnot require any line search procedure and provably converges to a weak solution at a rate of $O(epsilon-2/(p+1))$.
  arXiv  Detail & Related papers  (2022-05-06T13:29:14Z)
• A first-order primal-dual method with adaptivity to local smoothness [64.62056765216386]
  We consider the problem of finding a saddle point for the convex-concave objective $min_x max_y f(x) + langle Ax, yrangle - g*(y)$, where $f$ is a convex function with locally Lipschitz gradient and $g$ is convex and possibly non-smooth. We propose an adaptive version of the Condat-Vu algorithm, which alternates between primal gradient steps and dual steps.
  arXiv  Detail & Related papers  (2021-10-28T14:19:30Z)
• High-probability Bounds for Non-Convex Stochastic Optimization with Heavy Tails [55.561406656549686]
  We consider non- Hilbert optimization using first-order algorithms for which the gradient estimates may have tails. We show that a combination of gradient, momentum, and normalized gradient descent convergence to critical points in high-probability with best-known iteration for smooth losses.
  arXiv  Detail & Related papers  (2021-06-28T00:17:01Z)
• Almost sure convergence rates for Stochastic Gradient Descent and Stochastic Heavy Ball [17.33867778750777]
  We study gradient descent (SGD) and the heavy ball method (SHB) for the general approximation problem. For SGD, in the convex and smooth setting, we provide the first emphalmost sure convergence emphrates for a weighted average of the iterates.
  arXiv  Detail & Related papers  (2020-06-14T11:12:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.