U

Underfitting

• Category: Math & training
• What people say: The model cannot fit the training task well enough.
• What it actually means: A model or training setup has insufficient effective capacity, optimization, features, or training signal to capture useful patterns in the training data.
• In practice: Diagnose data and optimization first, then consider training longer, changing features, reducing excessive regularization, or increasing suitable capacity.
• Related terms: Overfitting, Loss Function, Hyperparameter