Life In 19x19

Alpha no Go - After the Match

Mirror: http://imgur.com/a/WefEN

mumps wrote:Aja Huang (5 dan) is the person from the DeepMind team who's playing the moves. I believe that is all he's doing, as apparently it's quite stressful enough just to do that.

I bet! Imagine having to sit for 5 hours long and having to promptly respond to every move within seconds.

Also: https://online-go.com/review/114161

Damn, Lee was playing so well at the beginning.

I want AlphaGo to win, but still pretty shocked.

Random comments:

- Too bad they didn't show the set up, the audience, etc.
- Not digging Lee Sedol's short haircut
- Would be very interesting if they give Lee Sedol food so we can watch him eat in stream
- Wanna see more of AlphaGo's screen and its confidence level

Subotai wrote:That is sad to hear. While video games are a great form of entertainment I think they have taken too much attention from chess/go.

Hopefully AlphaGo vs Lee Sedol will bring attention to the mass.

Not yet.

So far all we know are from the recent tweets: https://twitter.com/demishassabis

illluck wrote:Ah, yeah, the time limits don't seem likely to lead to super long games.

I hope Lee Sedol doesn't pick 6 hr long games...

Thanks pasky, I was also wondering some about the asynchronous evaluation, which you mentioned in the computer-go post. Would it be possible to have something like alpha-go on a 1-gpu machine, or is there no way to get it to scale down?

From the paper:

https://i.imgur.com/xxdWUtV ...