> [28] Question 2: Neural Networks Consider the function fi(21,22), shown in Table 2. 21 -1 -1 +1 +1 I 2 fi -1 +1+1 -1 -1 +1 -1 Table 2: Function fi(11, 12) (a) Draw a diagram in (I), Ia)-space to show the placement of a decision line (hyperplane) of a single Perceptron, that will correctly classify fi. (5) (b) Design a neural network consisting of a single neuron that uses a threshold activation function, that will correctly classify f1(21, #2) Show how you calculate the weights, and prove that the neuron correctly classifies f1. Draw a diagram of the network with all labels and values correctly indicated. (14) (c) The activation function of the BACKPROPAGATION algorithm is critical to suc- cessful learning. Write down functions for the sigmoid activation function and the tanh activation function, in terms of the neuron's net input value Discuss the similarities and differences between these activation functions in relation to their effect on network training. (9)