English
Related papers

Related papers: Sharp feature-learning transitions and Bayes-optim…

200 papers

When data is plentiful, the loss achieved by well-trained neural networks scales as a power-law $L \propto N^{-\alpha}$ in the number of network parameters $N$. This empirical scaling law holds for a wide variety of data modalities, and may…

Machine Learning · Computer Science 2020-04-24 Utkarsh Sharma , Jared Kaplan

Deep learning is also known as hierarchical learning, where the learner _learns_ to represent a complicated target function by decomposing it into a sequence of simpler functions to reduce sample and time complexity. This paper formally…

Machine Learning · Computer Science 2023-07-10 Zeyuan Allen-Zhu , Yuanzhi Li

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

Classifiers built with neural networks handle large-scale high dimensional data, such as facial images from computer vision, extremely well while traditional statistical methods often fail miserably. In this paper, we attempt to understand…

Machine Learning · Statistics 2020-02-04 Tianyang Hu , Zuofeng Shang , Guang Cheng

For three decades statistical mechanics has been providing a framework to analyse neural networks. However, the theoretically tractable models, e.g., perceptrons, random features models and kernel machines, or multi-index models and…

Machine Learning · Statistics 2025-06-02 Jean Barbier , Francesco Camilli , Minh-Toan Nguyen , Mauro Pastore , Rudy Skerk

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…

Machine Learning · Computer Science 2022-06-17 Jacob A. Zavatone-Veth , William L. Tong , Cengiz Pehlevan

Hard-label classification is usually trained with smooth surrogate losses, most prominently softmax cross-entropy. We isolate an asymptotic mechanism by which this mismatch between smooth surrogate and discrete labels produces power-law…

Machine Learning · Computer Science 2026-05-22 Marcel Kühn , Yoon Thelge , Bernd Rosenow

Large language models with a huge number of parameters, when trained on near internet-sized number of tokens, have been empirically shown to obey neural scaling laws: specifically, their performance behaves predictably as a power law in…

Machine Learning · Computer Science 2022-11-01 Alexander Maloney , Daniel A. Roberts , James Sully

In machine learning, the scaling law describes how the model performance improves with the model and data size scaling up. From a learning theory perspective, this class of results establishes upper and lower generalization bounds for a…

Machine Learning · Computer Science 2025-02-14 Shihong Ding , Haihan Zhang , Hanzhen Zhao , Cong Fang

The scaling law, a cornerstone of Large Language Model (LLM) development, predicts improvements in model performance with increasing computational resources. Yet, while empirically validated, its theoretical underpinnings remain poorly…

Machine Learning · Computer Science 2026-02-03 Chiwun Yang

We study the convergence of gradient methods for the training of mean-field single-hidden-layer neural networks with square loss. For this high-dimensional and non-convex optimization problem, most known convergence results are either…

Machine Learning · Computer Science 2025-07-22 Raphaël Barboni , Gabriel Peyré , François-Xavier Vialard

Feature learning is thought to be one of the fundamental reasons for the success of deep neural networks. It is rigorously known that in two-layer fully-connected neural networks under certain conditions, one step of gradient descent on the…

Machine Learning · Statistics 2025-04-11 Behrad Moniri , Donghwan Lee , Hamed Hassani , Edgar Dobriban

According to a popular viewpoint, neural networks learn from data by first identifying low-dimensional representations, and subsequently fitting the best model in this space. Recent works provide a formalization of this phenomenon when…

Machine Learning · Computer Science 2026-02-27 Andrea Montanari , Zihao Wang

Deep neural networks (DNN) with a huge number of adjustable parameters remain largely black boxes. To shed light on the hidden layers of DNN, we study supervised learning by a DNN of width $N$ and depth $L$ consisting of $NL$ perceptrons…

Disordered Systems and Neural Networks · Physics 2023-08-01 Hajime Yoshino

We perform an average case analysis of the generalization dynamics of large neural networks trained using gradient descent. We study the practically-relevant "high-dimensional" regime where the number of free parameters in the network is on…

Machine Learning · Statistics 2017-10-11 Madhu S. Advani , Andrew M. Saxe

Feature extraction - the ability to identify relevant properties of data - is a key factor underlying the success of deep learning. Yet, it has proved difficult to elucidate its nature within existing predictive theories, to the extent that…

Disordered Systems and Neural Networks · Physics 2025-08-29 Andrea Corti , Rosalba Pacelli , Pietro Rotondo , Marco Gherardi

Deep learning has recently revealed the existence of scaling laws, demonstrating that model performance follows predictable trends based on dataset and model sizes. Inspired by these findings and fascinating phenomena emerging in the…

Machine Learning · Statistics 2026-02-10 Mattia Rosso , Simone Rossi , Giulio Franzese , Markus Heinonen , Maurizio Filippone

Empirically, large-scale deep learning models often satisfy a neural scaling law: the test error of the trained model improves polynomially as the model size and data size grow. However, conventional wisdom suggests the test error consists…

Machine Learning · Computer Science 2025-06-11 Licong Lin , Jingfeng Wu , Sham M. Kakade , Peter L. Bartlett , Jason D. Lee

We consider the problem of learning a target function corresponding to a single hidden layer neural network, with a quadratic activation function after the first layer, and random weights. We consider the asymptotic limit where the input…

Machine Learning · Statistics 2025-02-10 Antoine Maillard , Emanuele Troiani , Simon Martin , Florent Krzakala , Lenka Zdeborová

We study the dynamics of optimization and the generalization properties of one-hidden layer neural networks with quadratic activation function in the over-parametrized regime where the layer width $m$ is larger than the input dimension $d$.…

Machine Learning · Computer Science 2021-03-22 Stefano Sarao Mannelli , Eric Vanden-Eijnden , Lenka Zdeborová