From the theory of SVD’s we know G can be decomposed as a sum of rank-many rankone matrices. Suppose that G is approximated by a rank-one matrix sqT with s ∈ Rn and q ∈ Rm with non-negative components. Can you use this fact to give a difficulty score or rating? What is the possible meaning of the vector s? Note one can use the top singular value decomposition to get this score vector!