cards/machine-learning/measurement/empirical-loss.md

To measure the generalization, we define a generalization error, $$ \begin{align} \mathcal G = …

The loss calculated on all the whole population