Benign Overfitting and Noisy Features

TL;DR

This paper investigates benign overfitting in random feature models, revealing that noise in features plays a crucial regularization role enabling models to generalize well despite overparameterization.

Contribution

It provides a new perspective on benign overfitting by highlighting the role of noise in features within random neural network models.

Findings

Benign overfitting occurs due to noise in features.

Noise in features acts as an implicit regularizer.

The phenomenon is analyzed in two-layer neural networks with fixed first layer weights.

Abstract

Modern machine learning often operates in the regime where the number of parameters is much higher than the number of data points, with zero training loss and yet good generalization, thereby contradicting the classical bias-variance trade-off. This \textit{benign overfitting} phenomenon has recently been characterized using so called \textit{double descent} curves where the risk undergoes another descent (in addition to the classical U-shaped learning curve when the number of parameters is small) as we increase the number of parameters beyond a certain threshold. In this paper, we examine the conditions under which \textit{Benign Overfitting} occurs in the random feature (RF) models, i.e. in a two-layer neural network with fixed first layer weights. We adopt a new view of random feature and show that \textit{benign overfitting} arises due to the noise which resides in such features (the noise may already be present in the data and propagate to the features or it may be added by the user to the features directly) and plays an important implicit regularization role in the phenomenon.