A priori estimates for classification problems using neural networks

We consider binary and multi-class classification problems using hypothesis classes of neural networks. For a given hypothesis class, we use Rademacher complexity estimates and direct approximation theorems to obtain a priori error estimates for regularized loss functionals.

stat.ML cs.LG math.NA Neural Networks and Applications Image and Signal Denoising Methods Model Reduction and Neural Networks Computer Science Mathematics

View graph Source

Paper

References (15)

01Representation formulas and pointwise properties for Barron functionsE. Weinan, Stephan Wojtowytsch2020 · Calculus of Variations and Partial Differential Equations · 104 citations In Library

02Towards a Mathematical Understanding of Neural Network-Based Machine Learning: what we know and what we don'tE. Weinan, Chao Ma, Stephan Wojtowytsch et al.2020 · CSIAM Transactions on Applied Mathematics · 151 citations In Library

03Kolmogorov width decay and poor approximators in machine learning: shallow neural networks, random feature models and neural tangent kernelsE. Weinan, Stephan Wojtowytsch2020 · Research in the Mathematical Sciences · 33 citations In Library

04Implicit Bias of Gradient Descent for Wide Two-layer Neural Networks Trained with the Logistic LossLénaïc Chizat, F. Bach2020 · Annual Conference Computational Learning Theory · 387 citations In Library

05Stability for the Training of Deep Neural Networks and Other ClassifiersL. Berlyand, P. Jabin, C. A. Safsten2020 · Mathematical Models and Methods in Applied Sciences · 10 citations In Library

06Barron Spaces and the Compositional Function Spaces for Neural Network ModelsE. Weinan, Chao Ma, Lei Wu2019 · arXiv.org · 117 citations In Library

07A priori estimates of the population risk for two-layer neural networksW. E, Chao Ma, Lei Wu2018 · Communications in Mathematical Sciences · 148 citations In Library

08A Vector-Contraction Inequality for Rademacher ComplexitiesAndreas Maurer2016 · International Conference on Algorithmic Learning Theory · 310 citations In Library

09Breaking the Curse of Dimensionality with Convex Neural NetworksF. Bach2014 · Journal of machine learning research · 821 citations In Library

10Modern Methods in the Calculus of Variations: L^p SpacesI. Fonseca, G. Leoni2007 · 626 citations

11Uniform Approximation by Neural NetworksY. Makovoz1998 · 113 citations

12Universal approximation bounds for superpositions of a sigmoidal functionA. Barron1993 · IEEE Transactions on Information Theory · 3.4k citations

Scroll for more · 3 remaining

Similar papers

Neural network approximation and estimation of classifiers with classification boundary in a Barron classAndrei Caragea, Philipp Petersen, Felix Voigtlaender2023 · The Annals of Applied Probability In Library

Neural network approximation and estimation of classifiers with classification boundary in a Barron classA. Caragea, P. Petersen, F. Voigtlaender2020 · The Annals of Applied Probability · 42 citations In Library

Nonlinear Weighted Directed Acyclic Graph and A Priori Estimates for Neural NetworksYuqing Li, Tao Luo, Chao Ma2021 · arXiv (Cornell University) · 2 citations In Library

Generalization and Estimation Error Bounds for Model-based Neural NetworksAvner Shultzman, Eyar Azar, Miguel R. D. Rodrigues et al.2023 · arXiv (Cornell University) · 13 citations In Library

Dimension-independent learning rates for high-dimensional classification problemsAndres Felipe Lerma-Pineda, Philipp Petersen, Simon Frieder et al.2025 · Analysis and Applications In Library

Adversarial Learning Guarantees for Linear Hypotheses and Neural NetworksPranjal Awasthi, Natalie S. Frank, Mehryar Mohri2020 · arXiv (Cornell University) · 66 citations In Library

Adversarial Learning Guarantees for Linear Hypotheses and Neural NetworksPranjal Awasthi, Natalie S. Frank, Mehryar Mohri2020 · International Conference on Machine Learning In Library

Nonlinear Weighted Directed Acyclic Graph and A Priori Estimates for Neural NetworksYuqing Li, Tao Luo, Chao Ma2022 · SIAM Journal on Mathematics of Data Science In Library

A Priori Estimates of the Population Risk for Two-layer Neural NetworksW. E, Chao Ma, Lei Wu2018 · Communications in Mathematical Sciences · 145 citations In Library

On Excess Risk Convergence Rates of Neural Network ClassifiersHyunouk Ko, Namjoon Suh, Xiaoming Huo2026 · IEEE Transactions on Information Theory In Library

Source of record

arXiv

44B

Welcome.

This panel sticks with you. Pick anything below and it opens right beside it, so you can dig through 60,000-plus records without ever losing your spot here.

Travel 44B

LibraryPapers, policy, standards, statute LabsEvery organization building AI ModelsIntelligence, price, and speed BenchmarksThe evaluation catalog IncidentsAI failures and harms SearchOne field across everything DashboardThe state of AI accountability in NY 44B RegistryThe Art. 44-B compliance portal