Quantifying Uncertainty for XGBRegressor

[WEILMAX]

Hello,

I came across your work when I was looking at a way to quantify the uncertainty of XGBRegressors. I really like your method of uncertainty quantification using virtual Ensembles as this is computationally very efficient and requires no additional memory. Would this work as well on XGBoost as on CatBoost and are there ways of constructing the XGBRegressor that fulfill the SGLB requirements?

Thanks again for the interesting work!