3 ms·
That can't be right. Amateurs do not beat pro players 25% of the time, yet single machine AlphaGo beats distributed AlphaGo 25% of the time.
by cjbprime 11y ago
That can't be right. Amateurs do not beat pro players 25% of the time, yet single machine AlphaGo beats distributed AlphaGo 25% of the time.
- simonh 11y agoThen it would seem that AlphaGo on a single machine isn't equivalent to an Amateur. Presumably it's really good even when running in a single machine, and the marginal improvement for each additional machine tails off quite quickly. But when you're pitching it at a top class opponent in a match getting global coverage, you want all the incremental improvement you can get.
- emcq 11y agoI don't think comparing win loss distributions is particularly insightful. By a single machine winning many games relative to the distributed version, really it's just saying that the value/policy network is more important than the monte carlo tree search. The main difference is the number of tree search evaluations you can do; it doesn't seem like they have a more sophisticated model in the parallel version. This suggests that there are systematic mistakes that the single 8 GPU machine makes compared to the distributed 280 GPU machine, but MCTS can smooth some of the individual mistakes over a bit. I would suspect that the general Go-playing population of humans do not share some of the systematic mistakes, so you likely won't be able to project these win/loss distributions to playing humans.