2. [35 marks] (Linear unit training) Manually train a linear unit based on the instances below using the gradient descent training rule. The initial values of weights are 0 0, ,-0, ,-0. The learning rate is 0.1 . Please update each weight at least three times. Algorithm: Gradient descent learning rule 1: Let D {(Xi, til 1, 2, , n} be the set of training instances. 2: Initialize the weight vector with random values, w 3:repeat 4 for each weight for each training instance (x, t) E D Compute the predicted output oj end for 11:end for 12 until stopping condition is met X1 0 X2 0 2.1 2. [35 marks] (Linear unit training) Manually train a linear unit based on the instances below using the gradient descent training rule. The initial values of weights are 0 0, ,-0, ,-0. The learning rate is 0.1 . Please update each weight at least three times. Algorithm: Gradient descent learning rule 1: Let D {(Xi, til 1, 2, , n} be the set of training instances. 2: Initialize the weight vector with random values, w 3:repeat 4 for each weight for each training instance (x, t) E D Compute the predicted output oj end for 11:end for 12 until stopping condition is met X1 0 X2 0 2.1