Twin Neural Network Regression

TL;DR

Twin neural network regression predicts differences between data points to efficiently create an ensemble of predictions, leading to improved accuracy and a built-in estimate of uncertainty, all with a single model.

Contribution

The paper introduces TNN regression, a novel method that leverages neural networks to predict differences, forming an ensemble without multiple models, enhancing accuracy and uncertainty estimation.

Findings

TNN regression competes with or outperforms state-of-the-art methods.

It intrinsically creates an ensemble of size twice the training set.

Violation of self-consistency provides uncertainty estimates.

Abstract

We introduce twin neural network (TNN) regression. This method predicts differences between the target values of two different data points rather than the targets themselves. The solution of a traditional regression problem is then obtained by averaging over an ensemble of all predicted differences between the targets of an unseen data point and all training data points. Whereas ensembles are normally costly to produce, TNN regression intrinsically creates an ensemble of predictions of twice the size of the training set while only training a single neural network. Since ensembles have been shown to be more accurate than single models this property naturally transfers to TNN regression. We show that TNNs are able to compete or yield more accurate predictions for different data sets, compared to other state-of-the-art methods. Furthermore, TNN regression is constrained by self-consistency conditions. We find that the violation of these conditions provides an estimate for the prediction uncertainty.