Popular gradient boosting method where the weak learners are decision trees.

These decision trees, as usual, are built iteratively based on a differentiable loss function. XGBoost is one of the most popular frameworks to support gradient boosted trees.