Answer:
Figure 1 shows the neural network that computes the XOR of two inputs.
Two Input unit: X₁ and X₂
One hidden layer with two hidden units: h₁ and h₂
Two Weights for hidden unit 1 = 20, 20
Bias for hidden unit 1 = -10
Two Weights for hidden unit 2 = -20, -20
Bias for hidden unit 2 = 30
Weights for output layer = 20, 20
Bias for output layer = -30
Activation function = sigmoid (σ)
Output at hidden unit and at the output unit Y is shown in Table attached in Fig 2