I want solution by hand not code pls Q$7.$ Given the data points below, train the neural network function, implemented in Q$1,$ using the backpropagation algorithm. To do so$,$ follow the steps below.

points$]$

$\backslash$table$[[-1\mathrm{.}5,0\mathrm{.}006],[-1,0\mathrm{.}29],[-0\mathrm{.}5,0\mathrm{.}617],[0,1\mathrm{.}0],[0\mathrm{.}5,1\mathrm{.}38],[1,1\mathrm{.}707],[1\mathrm{.}5,1\mathrm{.}92],[2,2\mathrm{.}0]]$

Randomly initialize the parameters: $W1,W2,$ b$1,$ and b$2.$

Randomly select a data point: $t$ and $p.$

Pass the parameters to backpropagate algorithm to get the updating weights and biases.

Check manually if the results you got are correct.

Repeat steps $2$ and $3,$ until there is no changes in weights and biases.

How many iterations did you do$?$ What are the converged values of Ws and bs$?$

Extra. How can you implement a neural network in a general way? In which you do not need to write another customized code for a $3-$layer, $4-$layer, or even more layers.