English
Related papers

Related papers: Weight Initialization and Variance Dynamics in Dee…

200 papers

In this paper, we provide the first precise distributional characterization of gradient descent iterates for general multi-layer neural networks under the canonical single-index regression model, in the `finite-width proportional regime'…

Machine Learning · Computer Science 2025-05-09 Qiyang Han , Masaaki Imaizumi

Neural network-based function approximation plays a pivotal role in the advancement of scientific computing and machine learning. Yet, training such models faces several challenges: (i) each target function often requires training a new…

Machine Learning · Computer Science 2025-10-13 Xinwen Hu , Yunqing Huang , Nianyu Yi , Peimeng Yin

This article derives and validates three principles for initialization and architecture selection in finite width graph neural networks (GNNs) with ReLU activations. First, we theoretically derive what is essentially the unique…

Machine Learning · Statistics 2023-06-21 Gage DeZoort , Boris Hanin

In this work, we generalize the ideas of Kaiming initialization to Graph Neural Networks (GNNs) and propose a new scheme (G-Init) that reduces oversmoothing, leading to very good results in node and graph classification tasks. GNNs are…

Machine Learning · Computer Science 2024-11-01 Dimitrios Kelesis , Dimitris Fotakis , Georgios Paliouras

We analyze the dynamics of training deep ReLU networks and their implications on generalization capability. Using a teacher-student setting, we discovered a novel relationship between the gradient received by hidden student nodes and the…

Machine Learning · Computer Science 2019-07-01 Yuandong Tian , Tina Jiang , Qucheng Gong , Ari Morcos

Training large transformer models from scratch for a target task requires lots of data and is computationally demanding. The usual practice of transfer learning overcomes this challenge by initializing the model with weights of a pretrained…

Fully-connected deep neural networks with weights initialized from independent Gaussian distributions can be tuned to criticality, which prevents the exponential growth or decay of signals propagating through the network. However, such…

Machine Learning · Computer Science 2024-06-13 Hannah Day , Yonatan Kahn , Daniel A. Roberts

Neural networks trained via gradient descent with random initialization and without any regularization enjoy good generalization performance in practice despite being highly overparametrized. A promising direction to explain this phenomenon…

Machine Learning · Computer Science 2022-05-17 Hancheng Min , Salma Tarmoun , Rene Vidal , Enrique Mallada

We explore the low-rank structure of the weight matrices in neural networks at the stationary points (limiting solutions of optimization algorithms) with $L2$ regularization (also known as weight decay). We show several properties of such…

Machine Learning · Computer Science 2025-08-21 Ilja Kuzborskij , Yasin Abbasi Yadkori

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…

Machine Learning · Computer Science 2023-12-29 Mingze Wang , Chao Ma

We study in this paper how to initialize the parameters of multinomial logistic regression (a fully connected layer followed with softmax and cross entropy loss), which is widely used in deep neural network (DNN) models for classification…

Computer Vision and Pattern Recognition · Computer Science 2018-09-18 Bowen Cheng , Rong Xiao , Yandong Guo , Yuxiao Hu , Jianfeng Wang , Lei Zhang

In recurrent neural networks, learning long-term dependency is the main difficulty due to the vanishing and exploding gradient problem. Many researchers are dedicated to solving this issue and they proposed many algorithms. Although these…

Machine Learning · Computer Science 2023-07-31 Ran Dou , Jose Principe

We propose a system for calculating a "scaling constant" for layers and weights of neural networks. We relate this scaling constant to two important quantities that relate to the optimizability of neural networks, and argue that a network…

Machine Learning · Computer Science 2022-05-16 Aaron Defazio , Léon Bottou

A recent line of work has established intriguing connections between the generalization/compression properties of a deep neural network (DNN) model and the so-called layer weights' stable ranks. Intuitively, the latter are indicators of the…

Machine Learning · Computer Science 2021-10-07 Bogdan Georgiev , Lukas Franken , Mayukh Mukherjee , Georgios Arvanitidis

Input-Convex Neural Networks (ICNNs) are networks that guarantee convexity in their input-output mapping. These networks have been successfully applied for energy-based modelling, optimal transport problems and learning invariances. The…

Machine Learning · Computer Science 2023-12-21 Pieter-Jan Hoedt , Günter Klambauer

With the prevalence of the Pretraining-Finetuning paradigm in transfer learning, the robustness of downstream tasks has become a critical concern. In this work, we delve into adversarial robustness in transfer learning and reveal the…

Computer Vision and Pattern Recognition · Computer Science 2024-04-02 Andong Hua , Jindong Gu , Zhiyu Xue , Nicholas Carlini , Eric Wong , Yao Qin

Inducing and leveraging sparse activations during training and inference is a promising avenue for improving the computational efficiency of deep networks, which is increasingly important as network sizes continue to grow and their…

Machine Learning · Computer Science 2024-02-27 Ilan Price , Nicholas Daultry Ball , Samuel C. H. Lam , Adam C. Jones , Jared Tanner

We provide a function space characterization of the inductive bias resulting from minimizing the $\ell_2$ norm of the weights in multi-channel convolutional neural networks with linear activations and empirically test our resulting…

Machine Learning · Computer Science 2022-07-12 Meena Jagadeesan , Ilya Razenshteyn , Suriya Gunasekar

In this paper, we present a novel approach to perform deep neural networks layer-wise weight initialization using Linear Discriminant Analysis (LDA). Typically, the weights of a deep neural network are initialized with: random values,…

Computer Vision and Pattern Recognition · Computer Science 2017-11-27 Michele Alberti , Mathias Seuret , Vinaychandran Pondenkandath , Rolf Ingold , Marcus Liwicki

This work, based on Random Matrix Theory (RMT), introduces a novel early-stopping strategy for Transformer training dynamics. Utilizing the Power Law (PL) fit to tansformer attention matrices as a probe, we demarcate training into three…

Machine Learning · Computer Science 2025-12-30 Jing He , Hua Jiang , Cheng Li , Siqian Xin , Shuzhen Yang
‹ Prev 1 4 5 6 7 8 10 Next ›