Hyperparameter Optimization

Gradient-Based Optimization

Approach that treats hyperparameters as optimizable parameters and calculates their gradient with respect to the model's loss. It enables efficient directional updates but requires differentiable objective functions.

← Tillbaka