Confidence Interval Construction and Conditional Variance Estimation with Dense ReLU Networks

TL;DR

This paper develops a residual-based framework for conditional variance estimation and confidence interval construction using dense ReLU networks, providing nonasymptotic bounds and a robust bootstrap method with theoretical guarantees.

Contribution

It introduces the first variance estimation bounds for ReLU networks and a robust bootstrap procedure for confidence intervals with coverage guarantees.

Findings

Nonasymptotic bounds for variance estimation under heteroscedastic noise.

First variance bounds for ReLU neural network estimators.

A robust bootstrap method with theoretical coverage guarantees.

Abstract

This paper addresses the problems of conditional variance estimation and confidence interval construction in nonparametric regression using dense networks with the Rectified Linear Unit (ReLU) activation function. We present a residual-based framework for conditional variance estimation, deriving nonasymptotic bounds for variance estimation under both heteroscedastic and homoscedastic settings. We relax the sub-Gaussian noise assumption, allowing the proposed bounds to accommodate sub-Exponential noise and beyond. Building on this, for a ReLU neural network estimator, we derive non-asymptotic bounds for both its conditional mean and variance estimation, representing the first result for variance estimation using ReLU networks. Furthermore, we develop a ReLU network based robust bootstrap procedure (Efron, 1992) for constructing confidence intervals for the true mean that comes with a theoretical guarantee on the coverage, providing a significant advancement in uncertainty quantification and the construction of reliable confidence intervals in deep learning settings.