CS MISC

I have edited out the questions from some years dealing with RNNs and CNNs. Therefore the marks of the questions may not add up to the total marks mentioned at the section headings. Section 1 [25] Short answer. No justification necessary. All questions in this section are worth equal marks. 1) True or False? A basic perceptron can classify points from only R2 or R3 . F 2) True or False? Two neural networks with identical architectures, trained with identical data, and producing identical satisfactory error rates must have identical parameter values. F 3) True or False? In back propagation, we try to minimize the parameters of a neural network. F 4) True or False? A basic perceptron with a restriction in its parameters, namely, w1 = 5, can classify two groups of linearly separable points, only if the clusters have a certain orientation relative to the coordinate axes. F 5) A basic perceptron takes its inputs from Rn . How many parameters would it have? n + 1 2 6) Consider 4 groups of points tightly clustered around 4 points in a st