What is the purpose of Batch Normalization in Deep Learning?

Answer choices

Select only one option

To reduce the number of trainable parameters in the network by taking an average of the weights across neurons within the same layer

To speed up the training process by optimizing the batch size hyperparameter.

To reduce the effects of variance within the same batch by normalizing the data fields across the batch

To add noise to the input data for preventing overfitting on input data