Uncertainty-aware Sensitivity Analysis Using R\'enyi Divergences

TL;DR

This paper introduces a Bayesian sensitivity analysis method that uses Rényi divergence to account for model uncertainty, improving the identification of important variables and interactions in nonlinear models.

Contribution

It extends derivative-based sensitivity analysis to a Bayesian framework by differentiating the Rényi divergence of predictive distributions, incorporating uncertainty.

Findings

Accurately identifies important variables and interactions.

Demonstrates improved reliability over alternative methods.

Validated on simulated and real datasets.

Abstract

For nonlinear supervised learning models, assessing the importance of predictor variables or their interactions is not straightforward because it can vary in the domain of the variables. Importance can be assessed locally with sensitivity analysis using general methods that rely on the model's predictions or their derivatives. In this work, we extend derivative based sensitivity analysis to a Bayesian setting by differentiating the R\'enyi divergence of a model's predictive distribution. By utilising the predictive distribution instead of a point prediction, the model uncertainty is taken into account in a principled way. Our empirical results on simulated and real data sets demonstrate accurate and reliable identification of important variables and interaction effects compared to alternative methods.