Clipped Gradient Methods for Nonsmooth Convex Optimization under Heavy-Tailed Noise: A Refined Analysis

• URL: http://arxiv.org/abs/2512.23178v1
• Date: Mon, 29 Dec 2025 03:35:53 GMT
• Title: Clipped Gradient Methods for Nonsmooth Convex Optimization under Heavy-Tailed Noise: A Refined Analysis
• Authors: Zijian Liu,
• Abstract summary: A simple yet effective operation, gradient clipping, is known to handle this new challenge successfully.<n>Our work improves upon the existing approach on two sides: better utilization of Freedman's inequality and finer bounds for clipping error under heavy-tailed noise.<n>To complement the study, we establish new lower bounds for both high-probability and in-expectation convergence.
• Score: 3.8357180714081327
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Optimization under heavy-tailed noise has become popular recently, since it better fits many modern machine learning tasks, as captured by empirical observations. Concretely, instead of a finite second moment on gradient noise, a bounded ${\frak p}$-th moment where ${\frak p}\in(1,2]$ has been recognized to be more realistic (say being upper bounded by $σ_{\frak l}^{\frak p}$ for some $σ_{\frak l}\ge0$). A simple yet effective operation, gradient clipping, is known to handle this new challenge successfully. Specifically, Clipped Stochastic Gradient Descent (Clipped SGD) guarantees a high-probability rate ${\cal O}(σ_{\frak l}\ln(1/δ)T^{1/{\frak p}-1})$ (resp. ${\cal O}(σ_{\frak l}^2\ln^2(1/δ)T^{2/{\frak p}-2})$) for nonsmooth convex (resp. strongly convex) problems, where $δ\in(0,1]$ is the failure probability and $T\in\mathbb{N}$ is the time horizon. In this work, we provide a refined analysis for Clipped SGD and offer two faster rates, ${\cal O}(σ_{\frak l}d_{\rm eff}^{-1/2{\frak p}}\ln^{1-1/{\frak p}}(1/δ)T^{1/{\frak p}-1})$ and ${\cal O}(σ_{\frak l}^2d_{\rm eff}^{-1/{\frak p}}\ln^{2-2/{\frak p}}(1/δ)T^{2/{\frak p}-2})$, than the aforementioned best results, where $d_{\rm eff}\ge1$ is a quantity we call the $\textit{generalized effective dimension}$. Our analysis improves upon the existing approach on two sides: better utilization of Freedman's inequality and finer bounds for clipping error under heavy-tailed noise. In addition, we extend the refined analysis to convergence in expectation and obtain new rates that break the known lower bounds. Lastly, to complement the study, we establish new lower bounds for both high-probability and in-expectation convergence. Notably, the in-expectation lower bounds match our new upper bounds, indicating the optimality of our refined analysis for convergence in expectation.

Related papers

• Convergence Rate for the Last Iterate of Stochastic Gradient Descent Schemes [0.0]
  We prove that SHB attains constant momentum with $F_*$ when Fdelta is convex.<n>We also prove that SHB attains constant momentum with $F_*$ when Fdelta is convex.
  arXiv  Detail & Related papers  (2025-07-09T20:59:23Z)
• Sign Operator for Coping with Heavy-Tailed Noise in Non-Convex Optimization: High Probability Bounds Under $(L_0, L_1)$-Smoothness [74.18546828528298]
  We show that SignSGD with Majority Voting can robustly work on the whole range of complexity with $kappakappakappakappa-1right, kappakappakappa-1right, kappakappakappa-1right, kappakappakappa-1right, kappakappakappa-1right, kappakappakappa-1right, kappakappakappa-1right, kappa
  arXiv  Detail & Related papers  (2025-02-11T19:54:11Z)
• 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)
• Breaking the Heavy-Tailed Noise Barrier in Stochastic Optimization Problems [56.86067111855056]
  We consider clipped optimization problems with heavy-tailed noise with structured density. We show that it is possible to get faster rates of convergence than $mathcalO(K-(alpha - 1)/alpha)$, when the gradients have finite moments of order. We prove that the resulting estimates have negligible bias and controllable variance.
  arXiv  Detail & Related papers  (2023-11-07T17:39:17Z)
• Breaking the Lower Bound with (Little) Structure: Acceleration in Non-Convex Stochastic Optimization with Heavy-Tailed Noise [28.780192812703948]
  We consider the optimization problem with smooth but not necessarily convex objectives in the heavy-tailed noise regime. We show that one can achieve a faster rate than that dictated by the lower bound $Omega(Tfrac1-p3p-2)$ with only tiny bit of structure. We also establish that it guarantees a high-probability convergence rate of $O(log(T/delta)Tfrac1-p3p-2)$ under a mild condition.
  arXiv  Detail & Related papers  (2023-02-14T00:23:42Z)
• Beyond Uniform Smoothness: A Stopped Analysis of Adaptive SGD [38.221784575853796]
  This work considers the problem of finding first-order stationary point of a non atau function with potentially constant smoothness using a gradient. We develop a technique that allows us to prove $mathcalO(fracmathrmpolylog(T)sigmatT)$ convergence rates without assuming uniform bounds on the noise.
  arXiv  Detail & Related papers  (2023-02-13T18:13:36Z)
• Private Stochastic Convex Optimization: Optimal Rates in $\ell_1$ Geometry [69.24618367447101]
  Up to logarithmic factors the optimal excess population loss of any $(varepsilon,delta)$-differently private is $sqrtlog(d)/n + sqrtd/varepsilon n.$ We show that when the loss functions satisfy additional smoothness assumptions, the excess loss is upper bounded (up to logarithmic factors) by $sqrtlog(d)/n + (log(d)/varepsilon n)2/3.
  arXiv  Detail & Related papers  (2021-03-02T06:53:44Z)
• Optimal Mean Estimation without a Variance [103.26777953032537]
  We study the problem of heavy-tailed mean estimation in settings where the variance of the data-generating distribution does not exist. We design an estimator which attains the smallest possible confidence interval as a function of $n,d,delta$.
  arXiv  Detail & Related papers  (2020-11-24T22:39:21Z)
• Accelerating Optimization and Reinforcement Learning with Quasi-Stochastic Approximation [2.294014185517203]
  This paper sets out to extend convergence theory to quasi-stochastic approximations. It is illustrated with applications to gradient-free optimization and policy gradient algorithms for reinforcement learning.
  arXiv  Detail & Related papers  (2020-09-30T04:44:45Z)
• Curse of Dimensionality on Randomized Smoothing for Certifiable Robustness [151.67113334248464]
  We show that extending the smoothing technique to defend against other attack models can be challenging. We present experimental results on CIFAR to validate our theory.
  arXiv  Detail & Related papers  (2020-02-08T22:02:14Z)

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.