Uncertainty-aware Sensitivity Analysis Using Rényi Divergences
This work addresses the challenge of uncertainty-aware sensitivity analysis for researchers and practitioners using Bayesian models, though it is incremental as it builds on existing derivative-based methods.
The authors tackled the problem of assessing variable importance in nonlinear supervised learning models by extending derivative-based sensitivity analysis to a Bayesian setting using Rényi divergences of predictive distributions, resulting in accurate and reliable identification of important variables and interactions compared to alternatives.
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 Rényi 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.