Re:AlphaGo Zero: Learning from scratch
by time-of-flight on Thursday October 19, 2017 @10:12AM
(#15508021)
Check the Nature paper (and I think this is one of the biggest take-aways from AlphaGo Zero):
"Finally, it uses a simpler tree search that relies upon this single neural network to evaluate positions and sample moves, without performing any Monte Carlo rollouts."
In this new version, MCTS is not even used to evaluate a position! Speaking as a Go player, the ability for the neural network to accurately evaluate a position without "reading" ahead is phenomenal (again, read the Nature paper last page for details).
"Finally, it uses a simpler tree search that relies upon this single neural network to evaluate positions and sample moves, without performing any Monte Carlo rollouts."
In this new version, MCTS is not even used to evaluate a position! Speaking as a Go player, the ability for the neural network to accurately evaluate a position without "reading" ahead is phenomenal (again, read the Nature paper last page for details).