Google’s DeepMind has been using applied reinforced learning techniques to the multiplication of mathematical matrices and can do it better than some 50 year-old human-made algorithms. Bored with ...