There's more...

A learning curve allows us to check whether the addition of training data leads to a benefit. It also allows us to estimate the contribution deriving from variance error and bias error. If the validation score and the training score converge with the size of the training set too low, we will not benefit from further training data.