Q$4$: Consider the following model

$f[x,]={}_3+w_3*cos[{}_2+w_2*exp[{}_1+w_1*sin[{}_0+w_0*x]]]$

with eight scalar parameters $=\{{}_0,w_0,{}_1,w_1,{}_2,w_2,{}_3,w_3\},$ and the following loss function

$l_i=(f[x_i,]-y_i{)}^2$

Where $x_i$ and $y_i$ represent the $i^{th}$ training input, and training output, respectively. Find the partial derivatives of the loss function with respect to all the parameters in $$ using backward differentiation.