We provide confidence intervals, prediction intervals, and hypothesis tests for variable importance for boosting with dropout and parallel training.