We derive a novel sensitivity analysis of input variables for predictive epistemic and aleatoric uncertainty. We use Bayesian neural networks with latent variables as a model class and illustrate the usefulness of our sensitivity analysis on real-world datasets. Our method increases the interpretability of complex black-box probabilistic models.
@article{arxiv.1712.03605,
title = {Sensitivity Analysis for Predictive Uncertainty in Bayesian Neural Networks},
author = {Stefan Depeweg and José Miguel Hernández-Lobato and Steffen Udluft and Thomas Runkler},
journal= {arXiv preprint arXiv:1712.03605},
year = {2017}
}