Batch Normalization ni the training process of AN Ns makes asignificant dif ference ot convergence rates because $($choose the most appropriate$)$:ni a somewhat deep ANN, at an intermediate weight layer fi a training sample $($mini$-$batch$)$ causes significant modulation of the weights, then the outputs from this layer wil ned ot be balanced by corresponding modulation of the succeeding weight layers down the line destabilizing convergence ni the training process; this si damped by normalizing the summed inputs into the activations at each layerb$)$ot reduce possibility of destabilizationni training, one uses very smal values of the learning rate parameter$)$ hte parameters of batch normalization are considered as additional parameters during training which increases the scope of optimizationd$)$does not realy help as the activation functions at alayer are ni any case bounded, while batch normalization si effected on the inputs ot these activation functions.