L2 Regularization
Domain: Evaluation
🎯 Motivation
Adds a penalty proportional to the square of weights. Encourages small, distributed weights. Also called weight decay.
📋 Concrete Example
L2 penalty: 0.01 * sum(weights^2). Large weights are penalized. The model uses many small weights instead of a few large ones. More robust, less overfitting.
🔗 Analogy
Packing a backpack - distribute weight evenly (L2) instead of putting everything in one pocket. Even distribution is more stable and robust.