Gradient boosting
Also: GBDT, XGBoost, LightGBM
Building an ensemble of small decision trees in sequence, each one fitted to the errors the ones before it left. Still the thing to beat on tabular data.
A definition is not an understanding. This is taught properly in How the main algorithms actually work, part of Supervised learning.
Learn this properlyTerms from the same topic