English
Related papers

Related papers: From Gradient Clipping to Normalization for Heavy …

200 papers

Recent studies have provided both empirical and theoretical evidence illustrating that heavy tails can emerge in stochastic gradient descent (SGD) in various scenarios. Such heavy tails potentially result in iterates with diverging…

Optimization and Control · Mathematics 2021-02-23 Hongjian Wang , Mert Gürbüzbalaban , Lingjiong Zhu , Umut Şimşekli , Murat A. Erdogdu

Recently, several studies consider the stochastic optimization problem but in a heavy-tailed noise regime, i.e., the difference between the stochastic gradient and the true gradient is assumed to have a finite $p$-th moment (say being upper…

Optimization and Control · Mathematics 2023-05-23 Zijian Liu , Zhengyuan Zhou

While stochastic gradient descent (SGD) is still the \emph{de facto} algorithm in deep learning, adaptive methods like Clipped SGD/Adam have been observed to outperform SGD across important tasks, such as attention models. The settings…

Optimization and Control · Mathematics 2020-10-26 Jingzhao Zhang , Sai Praneeth Karimireddy , Andreas Veit , Seungyeon Kim , Sashank J Reddi , Sanjiv Kumar , Suvrit Sra

In this work, we study the convergence \emph{in high probability} of clipped gradient methods when the noise distribution has heavy tails, ie., with bounded $p$th moments, for some $1<p\le2$. Prior works in this setting follow the same…

Optimization and Control · Mathematics 2023-04-05 Ta Duy Nguyen , Alina Ene , Huy L. Nguyen

The study of tail behaviour of SGD-induced processes has been attracting a lot of interest, due to offering strong guarantees with respect to individual runs of an algorithm. While many works provide high-probability guarantees, quantifying…

Machine Learning · Computer Science 2026-02-06 Aleksandar Armacki , Dragana Bajović , Dušan Jakovetić , Soummya Kar , Ali H. Sayed

The success of modern machine learning is due in part to the adaptive optimization methods that have been developed to deal with the difficulties of training large models over complex datasets. One such method is gradient clipping: a…

Machine Learning · Statistics 2024-10-08 Noah Marshall , Ke Liang Xiao , Atish Agarwala , Elliot Paquette

Modern machine learning is dominated by complex, overparameterized architectures capable of interpolating data and achieving zero training loss. For such models, we investigate the convergence properties of two popular modifications to…

Optimization and Control · Mathematics 2026-05-27 Aleksandr Lobanov , Anastasia Koloskova

Heavy-tailed noise is pervasive in modern machine learning applications, arising from data heterogeneity, outliers, and non-stationary stochastic environments. While second-order methods can significantly accelerate convergence in…

Optimization and Control · Mathematics 2025-10-14 Abdurakhmon Sadiev , Peter Richtárik , Ilyas Fatkhullin

Methods with adaptive stepsizes, such as AdaGrad and Adam, are essential for training modern Deep Learning models, especially Large Language Models. Typically, the noise in the stochastic gradients is heavy-tailed for the later ones.…

The empirical evidence indicates that stochastic optimization with heavy-tailed gradient noise is more appropriate to characterize the training of machine learning models than that with standard bounded gradient variance noise. Most…

Machine Learning · Computer Science 2026-01-28 Hongxu Chen , Ke Wei , Xiaoming Yuan , Luo Luo

By ensuring differential privacy in the learning algorithms, one can rigorously mitigate the risk of large models memorizing sensitive training data. In this paper, we study two algorithms for this purpose, i.e., DP-SGD and DP-NSGD, which…

Machine Learning · Computer Science 2022-06-28 Xiaodong Yang , Huishuai Zhang , Wei Chen , Tie-Yan Liu

Gradient clipping is commonly used in training deep neural networks partly due to its practicability in relieving the exploding gradient problem. Recently, \citet{zhang2019gradient} show that clipped (stochastic) Gradient Descent (GD)…

Machine Learning · Computer Science 2020-10-30 Bohang Zhang , Jikai Jin , Cong Fang , Liwei Wang

In this work we study high probability bounds for stochastic subgradient methods under heavy tailed noise. In this setting the noise is only assumed to have finite variance as opposed to a sub-Gaussian distribution for which it is known…

Optimization and Control · Mathematics 2024-04-16 Daniela A. Parletta , Andrea Paudice , Massimiliano Pontil , Saverio Salzo

Many tasks in modern machine learning are observed to involve heavy-tailed gradient noise during the optimization process. To manage this realistic and challenging setting, new mechanisms, such as gradient clipping and gradient…

Optimization and Control · Mathematics 2026-05-19 Zijian Liu

Gradient clipping is a standard training technique used in deep learning applications such as large-scale language modeling to mitigate exploding gradients. Recent experimental studies have demonstrated a fairly special behavior in the…

Machine Learning · Computer Science 2023-06-06 Amirhossein Reisizadeh , Haochuan Li , Subhro Das , Ali Jadbabaie

Neural network compression has been an increasingly important subject, not only due to its practical relevance, but also due to its theoretical implications, as there is an explicit connection between compressibility and generalization…

Machine Learning · Statistics 2024-02-13 Yijun Wan , Melih Barsbey , Abdellatif Zaidi , Umut Simsekli

Stochastic first-order methods such as Stochastic Extragradient (SEG) or Stochastic Gradient Descent-Ascent (SGDA) for solving smooth minimax problems and, more generally, variational inequality problems (VIP) have been gaining a lot of…

Optimization and Control · Mathematics 2022-11-02 Eduard Gorbunov , Marina Danilova , David Dobre , Pavel Dvurechensky , Alexander Gasnikov , Gauthier Gidel

Randomized subspace methods reduce per-iteration cost; however, in nonconvex optimization, most analyses are expectation-based, and high-probability bounds remain scarce even under sub-Gaussian noise. We first prove that randomized subspace…

Optimization and Control · Mathematics 2026-01-30 Gaku Omiya , Pierre-Louis Poirion , Akiko Takeda

Stochastic gradient algorithms are often unstable when applied to functions that do not have Lipschitz-continuous and/or bounded gradients. Gradient clipping is a simple and effective technique to stabilize the training process for problems…

Optimization and Control · Mathematics 2021-06-11 Vien V. Mai , Mikael Johansson

It has repeatedly been observed that loss minimization by stochastic gradient descent (SGD) leads to heavy-tailed distributions of neural network parameters. Here, we analyze a continuous diffusion approximation of SGD, called homogenized…

Machine Learning · Statistics 2024-02-05 Zhe Jiao , Martin Keller-Ressel