Risk Assessment for Machine Learning Models
Abstract
In this paper we propose a framework for assessing the risk associated with deploying a machine learning model in a specified environment. For that we carry over the risk definition from decision theory to machine learning. We develop and implement a method that allows to define deployment scenarios, test the machine learning model under the conditions specified in each scenario, and estimate the damage associated with the output of the machine learning model under test. Using the likelihood of each scenario together with the estimated damage we define \emph{key risk indicators} of a machine learning model. The definition of scenarios and weighting by their likelihood allows for standardized risk assessment in machine learning throughout multiple domains of application. In particular, in our framework, the robustness of a machine learning model to random input corruptions, distributional shifts caused by a changing environment, and adversarial perturbations can be assessed.
Cite
@article{arxiv.2011.04328,
title = {Risk Assessment for Machine Learning Models},
author = {Paul Schwerdtner and Florens Greßner and Nikhil Kapoor and Felix Assion and René Sass and Wiebke Günther and Fabian Hüger and Peter Schlicht},
journal= {arXiv preprint arXiv:2011.04328},
year = {2020}
}
Comments
8 pages, 5 figures, conference workshop