ngboost and XGBoostLSS

These are **competitors**: both extend gradient boosting frameworks to output full probability distributions rather than point predictions, offering alternative approaches to the same probabilistic prediction problem.