Consider a neural network $($NN$)$ with three layers. The input layer has two units. The

output layer consists of a single unit. The sole hidden layer consists of three units. We

use the sigmoid function as activation, on both second and third layers.

The network weights are given by the following weight matrices:

$])$

$1\mathrm{.}1,2\mathrm{.}4,-5\mathrm{.}2$

$([0\mathrm{.}6,3\mathrm{.}1,0\mathrm{.}4$

$($i$)$ Draw a picture representing the architecture of the network.

$($ii$)$ We feed the network the input vector $(x_1,x_2)=(0\mathrm{.}6,0\mathrm{.}2).$ Perform the forward

propagation and determine the values in the units of layers two and three.