Try to solve the same normal equation in problem 1.3 by using the spectral decomposition (numpy.linalg.svd) instead of pseudo-inverse. In your computation, assume that the inverse of the smallest eigenvalue ( zero) is an arbitrary large number. Is the solution the same as in problem 1.3? The pseudo-inverse operation contrasts with spectral decomposition, as it sets the inverse of the smallest eigenvalue to zero. Explain why pseudo-inverse and spectral decomposition yield the same solution