Batch Normalization is helpful because

Group of answer choices

It returns back the normalized mean and standard deviation of weights

None of these

It normalizes $($changes$)$ all the input before sending it to the next layer

It is a very efficient backpropagation technique