中文
相关论文

相关论文: Bad Global Minima Exist and SGD Can Reach Them

200 篇论文

Understanding the implicit bias of training algorithms is of crucial importance in order to explain the success of overparametrised neural networks. In this paper, we study the dynamics of stochastic gradient descent over diagonal linear…

机器学习 · 计算机科学 2021-12-08 Scott Pesme , Loucas Pillaud-Vivien , Nicolas Flammarion

Adaptive gradient methods such as Adam have gained increasing popularity in deep learning optimization. However, it has been observed that compared with (stochastic) gradient descent, Adam can converge to a different solution with a…

机器学习 · 计算机科学 2021-08-26 Difan Zou , Yuan Cao , Yuanzhi Li , Quanquan Gu

Multi-layer neural networks are among the most powerful models in machine learning, yet the fundamental reasons for this success defy mathematical understanding. Learning a neural network requires to optimize a non-convex high-dimensional…

机器学习 · 统计学 2022-06-08 Song Mei , Andrea Montanari , Phan-Minh Nguyen

Empirical studies show that gradient-based methods can learn deep neural networks (DNNs) with very good generalization performance in the over-parameterization regime, where DNNs can easily fit a random labeling of the training data. Very…

机器学习 · 计算机科学 2019-11-28 Yuan Cao , Quanquan Gu

Stochastic gradient descent (SGD) is a pillar of modern machine learning, serving as the go-to optimization algorithm for a diverse array of problems. While the empirical success of SGD is often attributed to its computational efficiency…

机器学习 · 统计学 2022-06-16 Courtney Paquette , Elliot Paquette , Ben Adlam , Jeffrey Pennington

The fundamental learning theory behind neural networks remains largely open. What classes of functions can neural networks actually learn? Why doesn't the trained network overfit when it is overparameterized? In this work, we prove that…

机器学习 · 计算机科学 2020-06-02 Zeyuan Allen-Zhu , Yuanzhi Li , Yingyu Liang

Stochastic gradient descent (SGD) has been found to be surprisingly effective in training a variety of deep neural networks. However, there is still a lack of understanding on how and why SGD can train these complex networks towards a…

机器学习 · 计算机科学 2019-01-03 Yi Zhou , Junjie Yang , Huishuai Zhang , Yingbin Liang , Vahid Tarokh

Local SGD is a communication-efficient variant of SGD for large-scale training, where multiple GPUs perform SGD independently and average the model parameters periodically. It has been recently observed that Local SGD can not only achieve…

机器学习 · 计算机科学 2023-03-10 Xinran Gu , Kaifeng Lyu , Longbo Huang , Sanjeev Arora

Graph Neural Networks (GNNs) have achieved notable success in learning from graph-structured data, owing to their ability to capture intricate dependencies and relationships between nodes. They excel in various applications, including…

机器学习 · 计算机科学 2023-11-29 Akansha A

We consider a one-hidden-layer leaky ReLU network of arbitrary width trained by stochastic gradient descent (SGD) following an arbitrary initialization. We prove that SGD produces neural networks that have classification accuracy…

机器学习 · 计算机科学 2021-02-16 Spencer Frei , Yuan Cao , Quanquan Gu

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…

机器学习 · 计算机科学 2024-05-30 Feng Chen , Daniel Kunin , Atsushi Yamamura , Surya Ganguli

The simplicity of gradient descent (GD) made it the default method for training ever-deeper and complex neural networks. Both loss functions and architectures are often explicitly tuned to be amenable to this basic local optimization. In…

机器学习 · 计算机科学 2019-04-30 Dmitrii Marin , Meng Tang , Ismail Ben Ayed , Yuri Boykov

Explaining the generalization characteristics of deep learning is an emerging topic in advanced machine learning. There are several unanswered questions about how learning under stochastic optimization really works and why certain…

机器学习 · 计算机科学 2022-04-01 Mahdi S. Hosseini , Mathieu Tuli , Konstantinos N. Plataniotis

Despite perfectly interpolating the training data, deep neural networks (DNNs) can often generalize fairly well, in part due to the "implicit regularization" induced by the learning algorithm. Nonetheless, various forms of regularization,…

机器学习 · 计算机科学 2022-02-23 Navid Azizan , Sahin Lale , Babak Hassibi

Neural network optimization remains one of the most consequential yet poorly understood challenges in modern AI research, where improvements in training algorithms can lead to enhanced feature learning in foundation models,…

机器学习 · 计算机科学 2025-12-23 Ansh Nagwekar

Proper regularization is critical for speeding up training, improving generalization performance, and learning compact models that are cost efficient. We propose and analyze regularized gradient descent algorithms for learning shallow…

机器学习 · 计算机科学 2018-06-08 Samet Oymak

Implicit deep learning has received increasing attention recently due to the fact that it generalizes the recursive prediction rules of many commonly used neural network architectures. Its prediction rule is provided implicitly based on the…

机器学习 · 计算机科学 2022-02-21 Tianxiang Gao , Hailiang Liu , Jia Liu , Hridesh Rajan , Hongyang Gao

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…

机器学习 · 计算机科学 2024-02-28 Hong T. M. Chu , Subhro Ghosh , Chi Thanh Lam , Soumendu Sundar Mukherjee

Neural networks are usually trained by some form of stochastic gradient descent (SGD)). A number of strategies are in common use intended to improve SGD optimization, such as learning rate schedules, momentum, and batching. These are…

神经与进化计算 · 计算机科学 2015-08-13 Thomas M. Breuel

This work is substituted by the paper in arXiv:2011.14066. Stochastic gradient descent is the de facto algorithm for training deep neural networks (DNNs). Despite its popularity, it still requires fine tuning in order to achieve its best…

机器学习 · 统计学 2020-12-02 Vatsal Shah , Anastasios Kyrillidis , Sujay Sanghavi