Question $2.[50$ points$]$

For this problem, change the definition of an M$-$P neuron so that both its inputs and outputs are

binary $(0$ or $1).$ View uv and wx each as two$-$bit binary $(00,01,10,$ or $11)$ numbers, and yz as the

$2$ low$-$order bits of the numerical addition of uv and wx $($see the figure below$).$

$($a$)$ Give weights and biases for an M$-$P network which generates z$.$

Because yz is the $2$ low$-$order bits of the numerical addition of uv and wx$,$ so z is the least

significant bit of the numerical addition of uv and wx$.$ And least significant bit of the binary

addition is essentially the result of an XOR operation, so we can use the following XOR gate as an

M$-$P network which generates z