We have a multilayer neural network with weights for the hidden layer: w$\_$hidden $=[(1,1,0\mathrm{.}5),(1,-1,1\mathrm{.}0)]$ and for the output layer: w$\_$out $=[1,1,-1]($both matrices include the biases$/$thresholds$).$ We have linear activation function and learning parameter eta $=0\mathrm{.}1.$ If we have inputs x $=[1,1,1].$T$,$ and the desired output d $=0.$ What is the local gradient $($delta$)$ of neuron $2$ that is in the hidden layer? Note: the neuron weights in the output layer are arranged like this... w$\_$output $=[$bias$,$ w$1,$ w$2].10$ points $2\mathrm{.}5-2\mathrm{.}53\mathrm{.}5-3\mathrm{.}5$