A recurrent network has a single hidden layer, $H=(h1$;$h2$;$h3),$ which uses an ReLU activation

function. At time $t,$ the activation levels of these three neurons are $(2,8,3),$ respectively. At time $t+1,$

their activations are $(4,1,6).$ Weights $w_{1,3}=0\mathrm{.}4$ and $w_{2,3}=0\mathrm{.}5$ are on the recurrent connections

from h$1$ to h$3,$ and h$2$ to h$3,$ respectively. Compute the following values:

a$.\frac{del{h}_{3,t+1}}{del{w}_{2,3}}=?$

b$.\frac{del{h}_{3,t+1}}{del{h}_{1,t}}=?$