But the new Student of Games algorithm gets around this limitation by combining guided search, self-play learning and game-theoretic reasoning, according to a new paper describing the algorithm ...