Liudmila Prokhorenkova: Uncertainty in Gradient Boosting via Ensembles

Data Fest Online 2020 Uncertainty Estimation in ML track Speaker: Liudmila Prokhorenkova, Yandex, HSE, MIPT In this video, I will show how to estimate predictive uncertainty in gradient boosting models. In particular, I will cover several methods for generating ensembles of such models and introduce a concept of a virtual ensemble that enables ensemble-based uncertainty estimation without any overhead in memory and time. I will also mention several tools for uncertainty estimation implemented in the CatBoost gradient boosting library. Register and get access to the tracks: Join the community: