The RMSprop optimizer is used for thebackpropagationstep during training. RMSprop keeps a moving average of the squared gradient for each weight and uses the moving average to divide the gradient. There are other hyperparameters such as the number ofhidden neurons, the depth of the network, the...