A Comparison of the Delta Method and the Bootstrap in Deep Learning Classification

07/04/2021

∙

by   Geir K. Nilsen, et al.

∙

0

∙

share

We validate the recently introduced deep learning classification adapted Delta method by a comparison with the classical Bootstrap. We show that there is a strong linear relationship between the quantified predictive epistemic uncertainty levels obtained from the two methods when applied on two LeNet-based neural network classifiers using the MNIST and CIFAR-10 datasets. Furthermore, we demonstrate that the Delta method offers a five times computation time reduction compared to the Bootstrap.