paper-with-me

Papers

Gradient Descent Converges Linearly to Flatter Minima than Gradient Flow in Shallow Linear Networks

2025-01-15 · Pierfrancesco Beneventano, Blake Woodworth

We study the gradient descent (GD) dynamics of a depth-2 linear neural network with a single input and output. We show that GD converges at an explicit linear rate to a global minimum of the training loss, even with a large stepsize -- about $2/\textrm{sharpness}$. It still converges for even larger stepsizes, but may do so very slowly. We also characterize the solution to which GD converges, which has lower norm and sharpness than the gradient flow solution. Our analysis reveals a trade off between the speed of convergence and the magnitude of implicit regularization. This sheds light on the benefits of training at the ``Edge of Stability'', which induces additional regularization by delaying convergence and may have implications for training more complex models.

📄 PDF Abstract BibTeX arXiv:2501.09137

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

SPEED The monocular depth estimation (MDE) is the task of estimating depth from a single frame. This information is an essential knowledge in many computer vision tasks such as scene…

Similar Papers 제목 키워드 기반

Sharpness-Aware Minimization Efficiently Selects Flatter Minima Late in Training

2024-10-14 · Zhanpeng Zhou, Mingze Wang, Yuchen Mao, Bingrui Li 외

Sharpness-Aware Minimization (SAM) has substantially improved the generalization of neural networks under various settings. Despite the success, its effectiveness remains poorly understood. In this work, we discover an i…

Implicit bias of deep linear networks in the large learning rate phase

2020-11-25 · Wei Huang, Weitao Du, Richard Yi Da Xu, Chunrui Liu

Most theoretical studies explaining the regularization effect in deep learning have only focused on gradient descent with a sufficient small learning rate or even gradient flow (infinitesimal learning rate). Such researc…

Binary Classification

Gradient Descent with Polyak's Momentum Finds Flatter Minima via Large Catapults

2023-11-25 · Prin Phunyaphibarn, Junghyun Lee, Bohan Wang, Huishuai Zhang 외

Although gradient descent with Polyak's momentum is widely used in modern machine and deep learning, a concrete understanding of its effects on the training trajectory remains elusive. In this work, we empirically show t…

When Will Gradient Methods Converge to Max-margin Classifier under ReLU Models?

2018-06-12 · ICLR 2019 5 · Tengyu Xu, Yi Zhou, Kaiyi Ji, Yingbin Liang

We study the implicit bias of gradient descent methods in solving a binary classification problem over a linearly separable dataset. The classifier is described by a nonlinear ReLU model and the objective function adopts…

Binary Classification

A Granger-Causal Perspective on Gradient Descent with Application to Pruning

2024-12-04 · Aditya Shah, Aditya Challa, Sravan Danda, Archana Mathur 외

Stochastic Gradient Descent (SGD) is the main approach to optimizing neural networks. Several generalization properties of deep networks, such as convergence to a flatter minima, are believed to arise from SGD. This arti…