Towards Algorithm-Agnostic Uncertainty Estimation: Predicting Classification Error in an Automated Machine Learning Setting

ICML 2020