The Slow Deterioration of the Generalization Error of the Random Feature Model

by   Chao Ma, et al.

The random feature model exhibits a kind of resonance behavior when the number of parameters is close to the training sample size. This behavior is characterized by the appearance of large generalization gap, and is due to the occurrence of very small eigenvalues for the associated Gram matrix. In this paper, we examine the dynamic behavior of the gradient descent algorithm in this regime. We show, both theoretically and experimentally, that there is a dynamic self-correction mechanism at work: The larger the eventual generalization gap, the slower it develops, both because of the small eigenvalues. This gives us ample time to stop the training process and obtain solutions with good generalization property.


page 1

page 2

page 3

page 4


Model, sample, and epoch-wise descents: exact solution of gradient flow in the random feature model

Recent evidence has shown the existence of a so-called double-descent an...

Train longer, generalize better: closing the generalization gap in large batch training of neural networks

Background: Deep learning models are typically trained using stochastic ...

Learning and Generalization in Overparameterized Neural Networks, Going Beyond Two Layers

Neural networks have great success in many machine learning applications...

High-dimensional dynamics of generalization error in neural networks

We perform an average case analysis of the generalization dynamics of la...

The Quenching-Activation Behavior of the Gradient Descent Dynamics for Two-layer Neural Network Models

A numerical and phenomenological study of the gradient descent (GD) algo...

A Reparameterization-Invariant Flatness Measure for Deep Neural Networks

The performance of deep neural networks is often attributed to their aut...