中文
相关论文

相关论文: Nonlinear Dynamics In Optimization Landscape of Sh…

200 篇论文

We study the loss surface of neural networks equipped with a hinge loss criterion and ReLU or leaky ReLU nonlinearities. Any such network defines a piecewise multilinear form in parameter space. By appealing to harmonic analysis we show…

机器学习 · 计算机科学 2018-07-24 Thomas Laurent , James von Brecht

In this paper, we analyze the landscape of the true loss of neural networks with one hidden layer and ReLU, leaky ReLU, or quadratic activation. In all three cases, we provide a complete classification of the critical points in the case…

机器学习 · 计算机科学 2022-07-07 Patrick Cheridito , Arnulf Jentzen , Florian Rossmannek

We prove that two-layer (Leaky)ReLU networks initialized by e.g. the widely used method proposed by He et al. (2015) and trained using gradient descent on a least-squares loss are not universally consistent. Specifically, we describe a…

机器学习 · 统计学 2022-06-10 David Holzmüller , Ingo Steinwart

We consider the optimization problem associated with fitting two-layer ReLU networks with $k$ hidden neurons, where labels are assumed to be generated by a (teacher) neural network. We leverage the rich symmetry exhibited by such models to…

机器学习 · 计算机科学 2021-09-22 Yossi Arjevani , Michael Field

We present a theoretical and empirical study of the gradient dynamics of overparameterized shallow ReLU networks with one-dimensional input, solving least-squares interpolation. We show that the gradient dynamics of such networks are…

机器学习 · 计算机科学 2019-06-20 Francis Williams , Matthew Trager , Claudio Silva , Daniele Panozzo , Denis Zorin , Joan Bruna

We investigate the training and generalization errors of overparameterized neural networks (NNs) with a wide class of leaky rectified linear unit (ReLU) functions. More specifically, we carefully upper bound both the convergence rate of the…

机器学习 · 计算机科学 2024-02-27 Yinglong Guo , Shaohan Li , Gilad Lerman

The training process of ReLU neural networks often exhibits complicated nonlinear phenomena. The nonlinearity of models and non-convexity of loss pose significant challenges for theoretical analysis. Therefore, most previous theoretical…

机器学习 · 计算机科学 2023-12-29 Mingze Wang , Chao Ma

The implicit bias towards solutions with favorable properties is believed to be a key reason why neural networks trained by gradient-based optimization can generalize well. While the implicit bias of gradient flow has been widely studied…

机器学习 · 计算机科学 2023-10-31 Yiwen Kou , Zixiang Chen , Quanquan Gu

The generalization mystery of overparametrized deep nets has motivated efforts to understand how gradient descent (GD) converges to low-loss solutions that generalize well. Real-life neural networks are initialized from small random values…

机器学习 · 计算机科学 2021-11-10 Kaifeng Lyu , Zhiyuan Li , Runzhe Wang , Sanjeev Arora

The implicit biases of gradient-based optimization algorithms are conjectured to be a major factor in the success of modern deep learning. In this work, we investigate the implicit bias of gradient flow and gradient descent in two-layer…

机器学习 · 计算机科学 2022-10-14 Spencer Frei , Gal Vardi , Peter L. Bartlett , Nathan Srebro , Wei Hu

We draw connections between simple neural networks and under-determined linear systems to comprehensively explore several interesting theoretical questions in the study of neural networks. First, we emphatically show that it is unsurprising…

数值分析 · 数学 2020-11-02 Austin R. Benson , Anil Damle , Alex Townsend

Neural networks with REctified Linear Unit (ReLU) activation functions (a.k.a. ReLU networks) have achieved great empirical success in various domains. Nonetheless, existing results for learning ReLU networks either pose assumptions on the…

机器学习 · 统计学 2019-05-01 Gang Wang , Georgios B. Giannakis , Jie Chen

We study the type of solutions to which stochastic gradient descent converges when used to train a single hidden-layer multivariate ReLU network with the quadratic loss. Our results are based on a dynamical stability analysis. In the…

机器学习 · 计算机科学 2023-07-03 Mor Shpigel Nacson , Rotem Mulayoff , Greg Ongie , Tomer Michaeli , Daniel Soudry

We prove that, for the fundamental regression task of learning a single neuron, training a one-hidden layer ReLU network of any width by gradient flow from a small initialisation converges to zero loss and is implicitly biased to minimise…

机器学习 · 计算机科学 2023-10-03 Dmitry Chistikov , Matthias Englert , Ranko Lazic

Despite remarkable performance on a variety of tasks, many properties of deep neural networks are not yet theoretically understood. One such mystery is the depth degeneracy phenomenon: the deeper you make your network, the closer your…

机器学习 · 统计学 2025-11-18 Cameron Jakub , Mihai Nica

We study the loss landscape of both shallow and deep, mildly overparameterized ReLU neural networks on a generic finite input dataset for the squared error loss. We show both by count and volume that most activation patterns correspond to…

机器学习 · 计算机科学 2024-02-09 Kedar Karhadkar , Michael Murray , Hanna Tseran , Guido Montúfar

In this paper, we study the implicit regularization of the gradient descent algorithm in homogeneous neural networks, including fully-connected and convolutional neural networks with ReLU or LeakyReLU activations. In particular, we study…

机器学习 · 计算机科学 2021-01-01 Kaifeng Lyu , Jian Li

The training of neural networks by gradient descent methods is a cornerstone of the deep learning revolution. Yet, despite some recent progress, a complete theory explaining its success is still missing. This article presents, for…

机器学习 · 统计学 2026-04-15 Etienne Boursier , Loucas Pillaud-Vivien , Nicolas Flammarion

This document proposes a parametric activation function (ac.f.) aimed at improving multidimensional nonlinear data regression. It is a established knowledge that nonlinear ac.f's are required for learning nonlinear datasets. This work shows…

机器学习 · 计算机科学 2025-10-03 Enda D. V. Bigarella

Dynamical systems theory has recently been applied in optimization to prove that gradient descent algorithms bypass so-called strict saddle points of the loss function. However, in many modern machine learning applications, the required…

机器学习 · 计算机科学 2024-09-12 Patrick Cheridito , Arnulf Jentzen , Florian Rossmannek
‹ 上一页 1 2 3 10 下一页 ›