中文
相关论文

相关论文: Is deeper better? It depends on locality of releva…

200 篇论文

Neural Tangent Kernel (NTK) theory is widely used to study the dynamics of infinitely-wide deep neural networks (DNNs) under gradient descent. But do the results for infinitely-wide networks give us hints about the behavior of real…

机器学习 · 计算机科学 2022-02-02 Mariia Seleznova , Gitta Kutyniok

A primary advantage of neural networks lies in their feature learning characteristics, which is challenging to theoretically analyze due to the complexity of their training dynamics. We propose a new paradigm for studying feature learning…

机器学习 · 计算机科学 2024-12-30 Haobo Zhang , Jianfa Lai , Yicheng Li , Qian Lin , Jun S. Liu

Several works have aimed to explain why overparameterized neural networks generalize well when trained by Stochastic Gradient Descent (SGD). The consensus explanation that has emerged credits the randomized nature of SGD for the bias of the…

机器学习 · 计算机科学 2021-02-24 Shengchao Liu , Dimitris Papailiopoulos , Dimitris Achlioptas

State-of-the-art, high capacity deep neural networks not only require large amounts of labelled training data, they are also highly susceptible to label errors in this data, typically resulting in large efforts and costs and therefore…

机器学习 · 计算机科学 2020-07-20 Christian Haase-Schütz , Rainer Stal , Heinz Hertlein , Bernhard Sick

Farquhar et al. [2021] show that correcting for active learning bias with underparameterised models leads to improved downstream performance. For overparameterised models such as NNs, however, correction leads either to decreased or…

机器学习 · 计算机科学 2021-12-15 Chelsea Murray , James U. Allingham , Javier Antorán , José Miguel Hernández-Lobato

Modular neural networks outperform nonmodular neural networks on tasks ranging from visual question answering to robotics. These performance improvements are thought to be due to modular networks' superior ability to model the compositional…

机器学习 · 计算机科学 2025-03-12 Akhilan Boopathy , Sunshine Jiang , William Yue , Jaedong Hwang , Abhiram Iyer , Ila Fiete

Deep neural networks (NN) have achieved great success in many applications. However, why do deep neural networks obtain good generalization at an over-parameterization regime is still unclear. To better understand deep NN, we establish the…

机器学习 · 统计学 2021-12-02 Yueming Lyu , Ivor Tsang

Understanding how feature learning affects generalization is among the foremost goals of modern deep learning theory. Here, we study how the ability to learn representations affects the generalization performance of a simple class of…

机器学习 · 计算机科学 2022-06-17 Jacob A. Zavatone-Veth , William L. Tong , Cengiz Pehlevan

In the context of supervised statistical learning, it is typically assumed that the training set comes from the same distribution that draws the test samples. When this is not the case, the behavior of the learned model is unpredictable and…

机器学习 · 计算机科学 2022-05-12 Antonio-Javier Gallego , Jorge Calvo-Zaragoza , Robert B. Fisher

Neural Tangent Kernel (NTK) is widely used to analyze overparametrized neural networks due to the famous result by Jacot et al. (2018): in the infinite-width limit, the NTK is deterministic and constant during training. However, this result…

机器学习 · 计算机科学 2022-07-22 Mariia Seleznova , Gitta Kutyniok

While neural networks are used for classification tasks across domains, a long-standing open problem in machine learning is determining whether neural networks trained using standard procedures are optimal for classification, i.e., whether…

机器学习 · 计算机科学 2023-05-03 Adityanarayanan Radhakrishnan , Mikhail Belkin , Caroline Uhler

A central theme of the modern machine learning paradigm is that larger neural networks achieve better performance on a variety of metrics. Theoretical analyses of these overparameterized models have recently centered around studying very…

机器学习 · 计算机科学 2024-10-10 Dhruva Karkada

We show how the success of deep learning could depend not only on mathematics but also on physics: although well-known mathematical theorems guarantee that neural networks can approximate arbitrary functions well, the class of functions of…

无序系统与神经网络 · 物理学 2017-09-13 Henry W. Lin , Max Tegmark , David Rolnick

Deep convolutional neural networks (CNNs) have demonstrated remarkable success in computer vision by supervisedly learning strong visual feature representations. However, training CNNs relies heavily on the availability of exhaustive…

计算机视觉与模式识别 · 计算机科学 2019-05-31 Jiabo Huang , Qi Dong , Shaogang Gong , Xiatian Zhu

Does over-parameterization eliminate sub-optimal local minima for neural networks? An affirmative answer was given by a classical result in [59] for 1-hidden-layer wide neural networks. A few recent works have extended the setting to…

机器学习 · 计算机科学 2020-11-17 Tian Ding , Dawei Li , Ruoyu Sun

Analysis of over-parameterized neural networks has drawn significant attention in recentyears. It was shown that such systems behave like convex systems under various restrictedsettings, such as for two-level neural networks, and when…

机器学习 · 计算机科学 2019-11-19 Cong Fang , Yihong Gu , Weizhong Zhang , Tong Zhang

The ability of learning useful features is one of the major advantages of neural networks. Although recent works show that neural network can operate in a neural tangent kernel (NTK) regime that does not allow feature learning, many works…

机器学习 · 计算机科学 2024-11-06 Mo Zhou , Rong Ge

For classification tasks, the performance of a deep neural network is determined by the structure of its decision boundary, whose geometry directly affects essential properties of the model, including accuracy and robustness. Motivated by a…

机器学习 · 计算机科学 2026-03-17 Matthew Burfitt , Jacek Brodzki , Pawel Dłotko

Conventional wisdom attributes the mysterious generalization abilities of overparameterized neural networks to gradient descent (and its variants). The recent volume hypothesis challenges this view: it posits that these generalization…

机器学习 · 计算机科学 2025-12-19 Yotam Alexander , Yonatan Slutzky , Yuval Ran-Milo , Nadav Cohen

We propose a new, more actionable view of neural network interpretability and data analysis by leveraging the remarkable matching effectiveness of representations derived from deep networks, guided by an approach for class-conditional…

计算与语言 · 计算机科学 2021-06-15 Allen Schmaltz