$1.$ Verify that the neural network $($with sigmoid activation functions$)$ in Figure $1$ approximates the not$$XOR$$ function by computing the values of the hidden units $($i$.$e$.,$ the outputs of the hidden nodes$)$ and the final output $\backslash ($ h$\_\{\backslash$theta$\}(\backslash$mathbf$\{$x$\})\backslash )$ for the given inputs below.

Table $1$: Input and output of not $($XOR$)$ function.

Figure $1$: Neural network that approximate not $($ XOR $)$ function.