English
Related papers

Related papers: Disentangling the Mechanisms Behind Implicit Regul…

200 papers

Understanding the implicit bias of Stochastic Gradient Descent (SGD) is one of the key challenges in deep learning, especially for overparametrized models, where the local minimizers of the loss function $L$ can form a manifold.…

Machine Learning · Computer Science 2022-07-29 Zhiyuan Li , Tianhao Wang , Sanjeev Arora

Modern machine learning paradigms, such as deep learning, occur in or close to the interpolation regime, wherein the number of model parameters is much larger than the number of data samples. In this work, we propose a regularity condition…

Machine Learning · Computer Science 2023-06-06 Chaoyue Liu , Dmitriy Drusvyatskiy , Mikhail Belkin , Damek Davis , Yi-An Ma

In this paper we aim to formally explain the phenomenon of fast convergence of SGD observed in modern machine learning. The key observation is that most modern learning architectures are over-parametrized and are trained to interpolate the…

Machine Learning · Computer Science 2018-06-18 Siyuan Ma , Raef Bassily , Mikhail Belkin

The phenomenon of implicit regularization has attracted interest in recent years as a fundamental aspect of the remarkable generalizing ability of neural networks. In a nutshell, it entails that gradient descent dynamics in many neural…

Machine Learning · Computer Science 2024-02-28 Hong T. M. Chu , Subhro Ghosh , Chi Thanh Lam , Soumendu Sundar Mukherjee

The success of denoising diffusion models raises important questions regarding their generalisation behaviour, particularly in high-dimensional settings. Notably, it has been shown that when training and sampling are performed perfectly,…

Machine Learning · Statistics 2025-07-08 Tyler Farghly , Patrick Rebeschini , George Deligiannidis , Arnaud Doucet

Stochastic gradient descent (SGD) has been the dominant optimization method for training deep neural networks due to its many desirable properties. One of the more remarkable and least understood quality of SGD is that it generalizes…

Machine Learning · Computer Science 2020-07-03 Erhan Bilal

When training neural networks, it has been widely observed that a large step size is essential in stochastic gradient descent (SGD) for obtaining superior models. However, the effect of large step sizes on the success of SGD is not well…

Machine Learning · Computer Science 2023-02-17 Amirkeivan Mohtashami , Martin Jaggi , Sebastian Stich

The success of deep learning has led to a rising interest in the generalization property of the stochastic gradient descent (SGD) method, and stability is one popular approach to study it. Existing works based on stability have studied…

Machine Learning · Statistics 2019-03-08 Yi Zhou , Yingbin Liang , Huishuai Zhang

Why does training deep neural networks using stochastic gradient descent (SGD) result in a generalization error that does not worsen with the number of parameters in the network? To answer this question, we advocate a notion of effective…

Machine Learning · Computer Science 2019-01-15 Vaishnavh Nagarajan , J. Zico Kolter

We propose a general yet simple theorem describing the convergence of SGD under the arbitrary sampling paradigm. Our theorem describes the convergence of an infinite array of variants of SGD, each of which is associated with a specific…

Machine Learning · Computer Science 2021-02-22 Robert Mansel Gower , Nicolas Loizou , Xun Qian , Alibek Sailanbayev , Egor Shulgin , Peter Richtarik

Neural networks typically generalize well when fitting the data perfectly, even though they are heavily overparameterized. Many factors have been pointed out as the reason for this phenomenon, including an implicit bias of stochastic…

Machine Learning · Computer Science 2025-02-04 Amit Peleg , Matthias Hein

We perform an experimental study of the dynamics of Stochastic Gradient Descent (SGD) in learning deep neural networks for several real and synthetic classification tasks. We show that in the initial epochs, almost all of the performance…

Machine Learning · Computer Science 2019-05-29 Preetum Nakkiran , Gal Kaplun , Dimitris Kalimeris , Tristan Yang , Benjamin L. Edelman , Fred Zhang , Boaz Barak

In this work, we reveal a strong implicit bias of stochastic gradient descent (SGD) that drives overly expressive networks to much simpler subnetworks, thereby dramatically reducing the number of independent parameters, and improving…

Machine Learning · Computer Science 2024-05-30 Feng Chen , Daniel Kunin , Atsushi Yamamura , Surya Ganguli

Previous work has examined the ability of larger capacity neural networks to generalize better than smaller ones, even without explicit regularizers, by analyzing gradient based algorithms such as GD and SGD. The presence of noise and its…

Machine Learning · Computer Science 2020-05-27 Arushi Gupta

Neural network has attracted great attention for a long time and many researchers are devoted to improve the effectiveness of neural network training algorithms. Though stochastic gradient descent (SGD) and other explicit gradient-based…

Optimization and Control · Mathematics 2020-02-11 Ren Liu , Xiaoqun Zhang

We study the generalization properties of stochastic gradient methods for learning with convex loss functions and linearly parameterized functions. We show that, in the absence of penalizations or constraints, the stability and…

Machine Learning · Computer Science 2016-05-27 Junhong Lin , Raffaello Camoriano , Lorenzo Rosasco

It is often observed that stochastic gradient descent (SGD) and its variants implicitly select a solution with good generalization performance; such implicit bias is often characterized in terms of the sharpness of the minima. Kleinberg et…

Machine Learning · Statistics 2024-05-28 Atsushi Nitanda , Ryuhei Kikuchi , Shugo Maeda , Denny Wu

The gradient noise of SGD is considered to play a central role in the observed strong generalization abilities of deep learning. While past studies confirm that the magnitude and the covariance structure of gradient noise are critical for…

Machine Learning · Computer Science 2020-06-22 Jingfeng Wu , Wenqing Hu , Haoyi Xiong , Jun Huan , Vladimir Braverman , Zhanxing Zhu

The early phase of training a deep neural network has a dramatic effect on the local curvature of the loss function. For instance, using a small learning rate does not guarantee stable optimization because the optimization trajectory has a…

LLM training is resource-intensive. Quantized training improves computational and memory efficiency but introduces quantization noise, which can hinder convergence and degrade model accuracy. Stochastic Rounding (SR) has emerged as a…

Machine Learning · Computer Science 2025-11-04 Taowen Liu , Marta Andronic , Deniz Gündüz , George A. Constantinides