At first thought, computing the similarity/distance between two datasets sounds easy, but in fact the problem is extremely difficult, explains Dr. James McCaffrey of Microsoft Research. A fairly ...