We have each data point x$=[[$ x$\_1$ x$\_2]]^$T in $^2,$ and we encounter the following data points: We are going to build a linear model $($i$.$e$.,$ using least square$)$ to predict the label of each data point x$=($x$\_1,$ x$\_2)$ as follows: $=$w$\_0+$w$\_1$ x$\_1+$w$\_2$ x$\_2$ where $($w$\_0,$ w$\_1,$ w$\_2)$ are unknown weights that we will learn based on the above training dataset. Here w$\_0$ is the intercept. $($a$)$ Gradient descent. Starting with $($w$\_0,$ w$\_1,$ w$\_2)=(0,0,0),$ we will run the gradient descent algorithm to learn these weights. Run for T$=4$ iterations and report the value of the weights and the corresponding total error at the end of each iteration. The learning rate is $\backslash$alpha $=0\mathrm{.}02.$ Note. For this question, you can compute each step manually or write a python program for it$.$ If you write a python program, you will need to submit the program together with your answers. $1($b$)$ Closed$-$form solution. Now we are going to use the closed$-$form solution to find $($w$\_0,$ w$\_1,$ w$\_2).$ Provide the values of $($w$\_0,$ w$\_1,$ w$\_2).$ What is the corresponding total error? Note. For this question, you are allowed to use any tool to perform matrix inverse or multiplication.