Regularization
The techniques that add additional constraints to the lost function to avoid excessive learning of the model.
Methods such as L1 and L2 regularization encourages the model to remain simpler and generalized by punishmenting large weight values.