Christopher I want to understand the concept of regularization in deep learning. How does it help reduce overfitting and improve model generalization? Can someone also explain common techniques like L1, L2, and dropout with simple examples?