Consider the feedforward neural network with the following training set: s (1) = h 1 1 1 i , with t (1) = h 1 1 i s (2) = h 1 1 −1 i , with t (2) = h 1 1 i s (3) = h −1 1 1 i , with t (3) = h −1 1 i s (4) = h −1 −1 1 i , with t (4) = h −1 1 i s (5) = h 1 −1 −1 i , with t (5) = h 1 −1 i s (6) = h 1 −1 1 i , with t (6) = h −1 −1 i thus s (1) are s (2) and in one class, s (3) and s (4) are in another class, s (5) is in a third class, and s (6) is in a fourth class. use the perceptron learning rule to try to find a set of weights and bias to correctly classify the training set. here we assume zero initial weights and bias, a learn rate α = 1, and a small but positive value of θ. 1. how many steps does it take for convergence? what is the final set of weight?