Glossary: Deep Learning Basics

Underfitting

When both training and validation loss stay high, because of too much regularization or a model that is too small.