Answer the question
In order to leave comments, you need to log in
Correct L2 regularization sequence?
The neural network began to retrain, regularization is required, I can’t understand at what stage this function is used.
At the stage of finding the gradient? New to this topic.
Answer the question
In order to leave comments, you need to log in
Regularization is applied in the backpropagation step.
In simple terms - you want to avoid very large weights for some neurons (aka overfitting). The more weight a certain neuron has, the more regularization for that neuron.
Picture from the Internet for clarity. I hope it helps to understand
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question