# How to compute variance of predicted values from a model that uses superlearner?

In addition to obtaining predicted values, is there a way to obtain the variance of predicted values from a model that uses superlearner? For instance, one may want to make inference about the difference under a different treatment.

SL.lib <- c("SL.glm", "SL.glmnet", "SL.xgboost", "SL.gam")
m.SL <- SuperLearner(Y = y_train, X = x_train,
cvControl = list(V = 20), SL.library = SL.lib)