Dropout Regularization Versus l2-Penalization in the Linear Model | Read Paper on Bytez