Maybe they can make a model that tells which set of rewarding experiments to do next!