Nonparametric Regression with Shallow Overparameterized Neural Networks Trained by GD with Early Stopping

Abstract

We explore the ability of overparameterized shallow neural networks to learn Lipschitz regression functions with and without label noise when trained by Gradient Descent (GD). To avoid the problem that in the presence of noisy labels, neural networks trained to nearly zero training error are inconsistent on this class, we propose an early stopping rule that allows us to show optimal rates. This provides an alternative to the result of Hu et al. (2021) who studied the performance of $\ell 2$ -regularized GD for training shallow networks in nonparametric regression which fully relied on the infinite-width network (Neural Tangent Kernel (NTK)) approximation. Here we present a simpler analysis which is based on a partitioning argument of the input space (as in the case of 1-nearest-neighbor rule) coupled with the fact that trained neural networks are smooth with respect to their inputs when trained by GD. In the noise-free case the proof does not rely on any kernelization and can be regarded as a finite-width result. In the case of label noise, by slightly modifying the proof, the noise is controlled using a technique of Yao, Rosasco, and Caponnetto (2007).

The report contains a critical issue. For details see the note: http://proceedings.mlr.press/v134/kuzborskij21a/kuzborskij21a.pdf

References in corpus (4)

• Simultaneous Model Selection and Optimization through Parameter-free Stochastic Learning
• Regularization Matters: A Nonparametric Perspective on Overparametrized Neural Network
• Over-parametrized deep neural networks do not generalize well
• Analysis of the rate of convergence of neural network regression estimates which are easy to implement