English
Related papers

Related papers: Implicit Acceleration and Feature Learning in Infi…

200 papers

Deep Neural Nets (DNNs) learn latent representations induced by their downstream task, objective function, and other parameters. The quality of the learned representations impacts the DNN's generalization ability and the coherence of the…

Machine Learning · Computer Science 2024-02-13 Nir Weingarten , Zohar Yakhini , Moshe Butman , Ran Gilad-Bachrach

Understanding how deep neural networks learn remains a fundamental challenge in modern machine learning. A growing body of evidence suggests that training dynamics undergo a distinct phase transition, yet our understanding of this…

Machine Learning · Computer Science 2025-05-21 Zhanpeng Zhou , Yongyi Yang , Mahito Sugiyama , Junchi Yan

While a lot of progress has been made in recent years, the dynamics of learning in deep nonlinear neural networks remain to this day largely misunderstood. In this work, we study the case of binary classification and prove various…

Machine Learning · Computer Science 2020-12-15 Remi Tachet , Mohammad Pezeshki , Samira Shabanian , Aaron Courville , Yoshua Bengio

Recent works have suggested that finite Bayesian neural networks may sometimes outperform their infinite cousins because finite networks can flexibly adapt their internal representations. However, our theoretical understanding of how the…

Machine Learning · Computer Science 2022-11-29 Jacob A. Zavatone-Veth , Abdulkadir Canatar , Benjamin S. Ruben , Cengiz Pehlevan

We theoretically characterize gradient descent dynamics in deep linear networks trained at large width from random initialization and on large quantities of random data. Our theory captures the ``wider is better" effect of…

Machine Learning · Computer Science 2025-06-17 Blake Bordelon , Cengiz Pehlevan

We study the optimization of wide neural networks (NNs) via gradient flow (GF) in setups that allow feature learning while admitting non-asymptotic global convergence guarantees. First, for wide shallow NNs under the mean-field scaling and…

Machine Learning · Computer Science 2022-04-25 Zhengdao Chen , Eric Vanden-Eijnden , Joan Bruna

The Information Bottleneck principle offers both a mechanism to explain how deep neural networks train and generalize, as well as a regularized objective with which to train models. However, multiple competing objectives are proposed in the…

Machine Learning · Computer Science 2021-01-06 Andreas Kirsch , Clare Lyle , Yarin Gal

Matrix completion problems arise in many applications including recommendation systems, computer vision, and genomics. Increasingly larger neural networks have been successful in many of these applications, but at considerable computational…

Machine Learning · Computer Science 2022-05-11 Adityanarayanan Radhakrishnan , George Stefanakis , Mikhail Belkin , Caroline Uhler

Understanding how convolutional neural networks (CNNs) can efficiently learn high-dimensional functions remains a fundamental challenge. A popular belief is that these models harness the local and hierarchical structure of natural data such…

Machine Learning · Statistics 2023-06-02 Francesco Cagnetta , Alessandro Favero , Matthieu Wyart

Neural networks are typically trained with a single learning rate across all layers. While recent empirical evidence suggests that assigning layer-specific learning rates can accelerate training, a principled understanding of the conditions…

Machine Learning · Computer Science 2026-05-26 Sihan Zeng , Sujay Bhatt , Sumitra Ganesh

Feature learning, or the ability of deep neural networks to automatically learn relevant features from raw data, underlies their exceptional capability to solve complex tasks. However, feature learning seems to be realized in different ways…

Machine Learning · Computer Science 2023-07-25 R. Aiudi , R. Pacelli , A. Vezzani , R. Burioni , P. Rotondo

In the past decade, deep neural networks have seen unparalleled improvements that continue to impact every aspect of today's society. With the development of high performance GPUs and the availability of vast amounts of data, learning…

Machine Learning · Computer Science 2021-05-12 Mohammad Ali Alomrani

Previous work has shown that DNNs with large depth $L$ and $L_{2}$-regularization are biased towards learning low-dimensional representations of the inputs, which can be interpreted as minimizing a notion of rank $R^{(0)}(f)$ of the learned…

Machine Learning · Computer Science 2024-08-16 Arthur Jacot

The choice of initial learning rate can have a profound effect on the performance of deep networks. We present a class of neural networks with solvable training dynamics, and confirm their predictions empirically in practical deep learning…

Machine Learning · Statistics 2020-03-05 Aitor Lewkowycz , Yasaman Bahri , Ethan Dyer , Jascha Sohl-Dickstein , Guy Gur-Ari

Deep neural networks perform well on classification tasks where data streams are i.i.d. and labeled data is abundant. Challenges emerge with non-stationary training data streams such as continual learning. One powerful approach that has…

While deep neural networks have demonstrated groundbreaking performance in various settings, these models often suffer from \emph{catastrophic forgetting} when trained on new tasks in sequence. Several works have empirically demonstrated…

Machine Learning · Computer Science 2024-06-21 Etash Guha , Vihan Lakshman

The prevailing thinking is that orthogonal weights are crucial to enforcing dynamical isometry and speeding up training. The increase in learning speed that results from orthogonal initialization in linear networks has been well-proven.…

Machine Learning · Computer Science 2021-07-22 Wei Huang , Weitao Du , Richard Yi Da Xu

Successful training of convolutional neural networks is often associated with sufficiently deep architectures composed of high amounts of features. These networks typically rely on a variety of regularization and pruning techniques to…

Computer Vision and Pattern Recognition · Computer Science 2017-10-23 Martin Mundt , Tobias Weis , Kishore Konda , Visvanathan Ramesh

We investigate the spectral properties of linear-width feed-forward neural networks, where the sample size is asymptotically proportional to network width. Empirically, we show that the spectra of weight in this high dimensional regime are…

Machine Learning · Computer Science 2023-11-09 Zhichao Wang , Andrew Engel , Anand Sarwate , Ioana Dumitriu , Tony Chiang

Existing analyses of neural network training often operate under the unrealistic assumption of an extremely small learning rate. This lies in stark contrast to practical wisdom and empirical studies, such as the work of J. Cohen et al.…

Machine Learning · Computer Science 2023-10-20 Kwangjun Ahn , Sébastien Bubeck , Sinho Chewi , Yin Tat Lee , Felipe Suarez , Yi Zhang