Consider the following MLP network. A new input pattern is presented to the network and training proceeds as follows. The actual outputs of the network are given by $z=([0\mathrm{.}32,0\mathrm{.}05,0\mathrm{.}67]{)}^T$ and the corresponding target values are given by $t=([1,1,1]{)}^T.$ The weights are also shown in the figure.

a$)$ What is the error $($local gradient$)$ for each of the output units?

b$)$ What is the error for each of the hidden unit$-2$ given that its activation fort he pattern being processed is currently $y_2=0\mathrm{.}5?$