The front page currently has a post on Deepmind's latest AI which can solve International Math Olympiad problems at a very high level. They state that this is in part built upon the AlphaZero reinforcement learning algorithm. I was curious as to any good resources to learn about this, as it appears that the AlphaZero implementation was never made public.
Good Tutorials on the Implementation of AlphaZero | Better HN
