English
Related papers

Related papers: The Expressive Power of Tuning Only the Normalizat…

200 papers

A recent line of research on deep learning focuses on the extremely over-parameterized setting, and shows that when the network width is larger than a high degree polynomial of the training sample size $n$ and the inverse of the target…

Machine Learning · Computer Science 2022-01-03 Zixiang Chen , Yuan Cao , Difan Zou , Quanquan Gu

Despite extensive studies, the underlying reason as to why overparameterized neural networks can generalize remains elusive. Existing theory shows that common stochastic optimizers prefer flatter minimizers of the training loss, and thus a…

Machine Learning · Computer Science 2023-07-25 Kaiyue Wen , Zhiyuan Li , Tengyu Ma

Handling the ever-increasing scale of contemporary deep learning and transformer-based models poses a significant challenge. Overparameterized Transformer networks outperform prior art in Natural Language processing and Computer Vision.…

Machine Learning · Computer Science 2024-11-05 Soheil Zibakhsh Shabgahi , Mohammad Sohail Shariff , Farinaz Koushanfar

Deep artificial neural networks achieve surprising generalization abilities that remain poorly understood. In this paper, we present a new approach to analyzing generalization for deep feed-forward ReLU networks that takes advantage of the…

Machine Learning · Computer Science 2023-07-06 Ramchandran Muthukumar , Jeremias Sulam

It is well-known that modern neural networks are vulnerable to adversarial examples. To mitigate this problem, a series of robust learning algorithms have been proposed. However, although the robust training error can be near zero via some…

Machine Learning · Computer Science 2022-10-17 Binghui Li , Jikai Jin , Han Zhong , John E. Hopcroft , Liwei Wang

Normalization methods play an important role in enhancing the performance of deep learning while their theoretical understandings have been limited. To theoretically elucidate the effectiveness of normalization, we quantify the geometry of…

Machine Learning · Statistics 2019-10-29 Ryo Karakida , Shotaro Akaho , Shun-ichi Amari

Reinforcement learning (RL) is a key post-pretraining step for aligning large language models (LLMs) with complex tasks and human preferences. While it is often assumed that RL fine-tuning requires updating most of a model's parameters, we…

Machine Learning · Computer Science 2025-07-30 Andrii Balashov

The success of deep networks has been attributed in part to their expressivity: per parameter, deep networks can approximate a richer class of functions than shallow networks. In ReLU networks, the number of activation patterns is one…

Machine Learning · Statistics 2019-10-22 Boris Hanin , David Rolnick

ReLU neural networks define piecewise linear functions of their inputs. However, initializing and training a neural network is very different from fitting a linear spline. In this paper, we expand empirically upon previous theoretical work…

Machine Learning · Statistics 2016-11-30 Kevin K. Chen , Anthony Gamst , Alden Walker

When approaching a novel visual recognition problem in a specialized image domain, a common strategy is to start with a pre-trained deep neural network and fine-tune it to the specialized domain. If the target domain covers a smaller visual…

Computer Vision and Pattern Recognition · Computer Science 2017-07-31 Frederick Tung , Srikanth Muralidharan , Greg Mori

Transformer networks have achieved remarkable empirical success across a wide range of applications, yet their theoretical expressive power remains insufficiently understood. In this paper, we study the expressive capabilities of…

Machine Learning · Computer Science 2026-03-04 Linyan Gu , Lihua Yang , Feng Zhou

Neural networks are regularly employed in adaptive control of nonlinear systems and related methods of reinforcement learning. A common architecture uses a neural network with a single hidden layer (i.e. a shallow network), in which the…

Optimization and Control · Mathematics 2024-04-18 Andrew Lamperski , Tyler Lekang

We introduce a novel stochastic regularization technique for deep neural networks, which decomposes a layer into multiple branches with different parameters and merges stochastically sampled combinations of the outputs from the branches…

Machine Learning · Computer Science 2019-10-04 Wonpyo Park , Paul Hongsuck Seo , Bohyung Han , Minsu Cho

With the advancement of deep learning, reducing computational complexity and memory consumption has become a critical challenge, and ternary neural networks (NNs) that restrict parameters to $\{-1, 0, +1\}$ have attracted attention as a…

Machine Learning · Computer Science 2026-04-28 Yuta Nakahara , Manabu Kobayashi , Toshiyasu Matsushima

We investigate filter level sparsity that emerges in convolutional neural networks (CNNs) which employ Batch Normalization and ReLU activation, and are trained with adaptive gradient descent techniques and L2 regularization or weight decay.…

Machine Learning · Computer Science 2019-04-08 Dushyant Mehta , Kwang In Kim , Christian Theobalt

We develop exact representations of training two-layer neural networks with rectified linear units (ReLUs) in terms of a single convex program with number of variables polynomial in the number of training samples and the number of hidden…

Machine Learning · Computer Science 2020-08-18 Mert Pilanci , Tolga Ergen

Transfer learning from ImageNet is the go-to approach when applying deep learning to medical images. The approach is either to fine-tune a pre-trained model or use it as a feature extractor. Most modern architecture contain batch…

Computer Vision and Pattern Recognition · Computer Science 2021-02-11 Fahdi Kanavati , Masayuki Tsuneki

This study explores the number of neurons required for a Rectified Linear Unit (ReLU) neural network to approximate multivariate monomials. We establish an exponential lower bound on the complexity of any shallow network approximating the…

Machine Learning · Computer Science 2023-05-17 Itai Shapira

In this paper we investigate the family of functions representable by deep neural networks (DNN) with rectified linear units (ReLU). We give an algorithm to train a ReLU DNN with one hidden layer to *global optimality* with runtime…

Machine Learning · Computer Science 2018-03-01 Raman Arora , Amitabh Basu , Poorya Mianjy , Anirbit Mukherjee

A practical limitation of deep neural networks is their high degree of specialization to a single task and visual domain. Recently, inspired by the successes of transfer learning, several authors have proposed to learn instead universal,…

Computer Vision and Pattern Recognition · Computer Science 2018-03-28 Sylvestre-Alvise Rebuffi , Hakan Bilen , Andrea Vedaldi