Go has fallen to computer domination?

[towforce]

I know the incentive isn't there because "we've already beaten the World Human Champion", but do you think a project like this could work for chess?

A Google Deepmind employee has already done it - link.

[Werewolf]

I know the incentive isn't there because "we've already beaten the World Human Champion", but do you think a project like this could work for chess?

A Google Deepmind employee has already done it - link.

Yes I know, I use Giraffe quite a bit.

What I meant was: a project on the same scale using lots of GPUs.
GPUs don't seem to be of much use in chess, but that might be because of the nature of old conventional programs.

Sorry that wasn't at all clear from my earlier post!!

[Milos]

can the software be converted to chess use and do you think it will be comparable to stockfish/komodo?

It could but it would bring absolutely nothing as it has been the case with previous N tries (N being a large 2 digit number).
8 different piece types are something no deep learning algorithm would (probably ever) be able to cope (19x19 binary picture is far simpler than 8x8x8 3D "object").

[Daniel Shawul]

Wow this is amazing news. What is interesting the approach seems to have a potential to contribute to general AI, unlike the brute-force method used by DeepBlue!
Let this be a lesson to all the alpha-beta optimizers in here (points finger away from self )

[lucasart]

Fan Hui is 2p professional. 5-0 suggests at least 4p for AlphaGo. Also, it beats normal MC programs like Crazy Stone (say 5 dan on i7) by 99.8%, which suggests 8-10 more stones, also above 4p. This thing runs on 170 GPU cards and 1,200 standard processors, and assuming quadrupling for a stone, it has 5 stones advantage over normal hardware on KGS (say a cluster of 24 regular cores). Lee Sedol is above regular 9p. Let's see.

I would bet on Alpha Go. Lee Sedol is toast.

[EroSennin]

You should also know the same human won 2 out of 5 unofficial blitz games.

[whereagles]

I gave the DeepMind link to a friend of mine who has some experience in neuron networks & stuff and has made a chess engine like 25 years ago (lol)...

He was pretty impressed.. GO was a personal favourite of him and this got him aching to get back to mind games programming

[duncan]

You should also know the same human won 2 out of 5 unofficial blitz games.

do you have a link to this?

[Laskos]

Fan Hui is 2p professional. 5-0 suggests at least 4p for AlphaGo. Also, it beats normal MC programs like Crazy Stone (say 5 dan on i7) by 99.8%, which suggests 8-10 more stones, also above 4p. This thing runs on 170 GPU cards and 1,200 standard processors, and assuming quadrupling for a stone, it has 5 stones advantage over normal hardware on KGS (say a cluster of 24 regular cores). Lee Sedol is above regular 9p. Let's see.

I would bet on Alpha Go. Lee Sedol is toast.

Hard to say, my extrapolation seems to give some good 9p for cluster bot, but Lee Sedol is not a regular 9p. In their paper, using normal Elo points, they give 3200 to the cluster bot, 3500 to Sedol, and Sedol would be clear favorite. But I don't like their Elo system used in the context of Go, they assume that the probability of 7d beating 6d is the same as 8k beating 9k, which is wrong by a lot.

If these commenting Chinese super-champions aren't simply showing off for humanity, I think their opinion that the bot is still weak compared to super-top is worth considering.

[Isaac]

You should also know the same human won 2 out of 5 unofficial blitz games.

do you have a link to this?

It's in the paper.

By the way they played in October, now of course Alphago has been improved.
I'd bet on the bot.