Nice introductory post! Being a go player what did you think of Google DeepMind's performance against the human player?
You are viewing a single comment's thread from:
Nice introductory post! Being a go player what did you think of Google DeepMind's performance against the human player?
Thank you!
I think Lee Sedol could have played a bit better, and it's quite obvious there are still some oddities in DeepMind's style of play. It seems to not be able to calculate a proper tesuji whenever it is playing a losing game. The thing is that because it's programmed to always play the best move, its programming seems to lack the imagination of a human player, thus it ends up playing a lot of clearly bad moves whenever it can't think of a one "best move".
Whenever it has gained the advantage though, that's a whole another thing.
I ended up losing 1 BTC because I made a bet against DeepMind.
I'm thinking Ke Jie might give it a run for it's life if they ever get to play, although DeepMind is quite strong when it plays correctly, and human is prone to err too.