Google's AlphaGo team has been working on chess

Discussion of chess software programming and technical issues.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
pkappler
Posts: 38
Joined: Thu Mar 09, 2006 1:19 am

Google's AlphaGo team has been working on chess

Post by pkappler » Wed Dec 06, 2017 4:10 am


Modern Times
Posts: 2142
Joined: Thu Jun 07, 2012 9:02 pm

Re: Google's AlphaGo team has been working on chess

Post by Modern Times » Wed Dec 06, 2017 6:18 am

Incredible:
In chess, AlphaZero outperformed Stockfish after just 4 hours
.

Opinions expressed here are my own, and not necessarily those of the CCRL Group.

Xann
Posts: 125
Joined: Sat Jan 22, 2011 6:14 pm
Location: Lille, France

Re: Google's AlphaGo team has been working on chess

Post by Xann » Wed Dec 06, 2017 6:31 am

Modern Times wrote:Incredible:
In chess, AlphaZero outperformed Stockfish after just 4 hours
Time is misleading in DeepMind's papers, as they use thousands of "computers" (not even commercially available). Money would be a better measure.

xcombelle
Posts: 2
Joined: Tue Aug 25, 2015 4:05 pm

Re: Google's AlphaGo team has been working on chess

Post by xcombelle » Wed Dec 06, 2017 8:21 am

Money would be a better measure.
The AlphaZero training system costed $ 4 millions of hardware. (figures given for alpha go zero, don't have source under hand)

smatovic
Posts: 479
Joined: Wed Mar 10, 2010 9:18 pm
Location: Germany
Contact:

Re: Google's AlphaGo team has been working on chess

Post by smatovic » Wed Dec 06, 2017 8:22 am

Money would be a better measure.
Or maybe games for training.....

NeuroChess 120 000
Giraffe (est.): 10 000 000
AlphaZero Chess: 44 000 000

--
Srdja

sasachess
Posts: 24
Joined: Wed Nov 05, 2014 10:28 am
Location: Italy
Contact:

Re: Google's AlphaGo team has been working on chess

Post by sasachess » Wed Dec 06, 2017 8:32 am

Evaluation speed:
AlphaZero 80K
Stockfish 70.000K

What?! :shock:

Fulvio
Posts: 91
Joined: Fri Aug 12, 2016 6:43 pm

Re: Google's AlphaGo team has been working on chess

Post by Fulvio » Wed Dec 06, 2017 9:31 am

pkappler wrote:Today is a big day in computer chess:

https://arxiv.org/abs/1712.01815
https://arxiv.org/pdf/1712.01815.pdf
"Instead of a handcrafted evaluation function and move ordering heuristics, AlphaZero utilises a deep neural network (p,v) = fθ(s) with parameters θ.
This neural network takes the board position s as an input and outputs a vector of move probabilities p with components pa = Pr(a|s) for each action a, and a scalar value v estimating the expected outcome z from position s"

This seems normal to me.

"Instead of an alpha-beta search with domain-specific enhancements, AlphaZero uses a general-purpose Monte-Carlo tree search (MCTS) algorithm. Each search consists of a series of simulated games of self-play that traverse a tree from root to leaf. Each simulation proceeds by selecting in each state a move with low visit count, high move probability and high value" [emphasis mine]

This is interesting. If I understand it correctly, it basically goes deeper only after reaching a high level of hash table hits.


"AlphaZero vs Stockfish: 25 win for AlphaZero, 25 draw, 0 loss (each program was given 1 minute of thinking time per move, strongest skill level using 64 threads and a hash size of 1GB)"

This is sci-fi. I do not have a 64 core machine but on my pc Stockfish do not sacrifice a Knight for 2 pawns:
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.d3 Bc5 5.Bxc6 dxc6 6.O-O Nd7 7.Nbd2 O-O 8.Qe1 f6 9.Nc4 Rf7 10.a4 Bf8 11.Kh1 Nc5 12.a5 Ne6 13.Ncxe5?

User avatar
Guenther
Posts: 2273
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Contact:

Re: Google's AlphaGo team has been working on chess

Post by Guenther » Wed Dec 06, 2017 10:03 am

Fulvio wrote:
pkappler wrote:Today is a big day in computer chess:

https://arxiv.org/abs/1712.01815
https://arxiv.org/pdf/1712.01815.pdf
"Instead of a handcrafted evaluation function and move ordering heuristics, AlphaZero utilises a deep neural network (p,v) = fθ(s) with parameters θ.
This neural network takes the board position s as an input and outputs a vector of move probabilities p with components pa = Pr(a|s) for each action a, and a scalar value v estimating the expected outcome z from position s"

This seems normal to me.

"Instead of an alpha-beta search with domain-specific enhancements, AlphaZero uses a general-purpose Monte-Carlo tree search (MCTS) algorithm. Each search consists of a series of simulated games of self-play that traverse a tree from root to leaf. Each simulation proceeds by selecting in each state a move with low visit count, high move probability and high value" [emphasis mine]

This is interesting. If I understand it correctly, it basically goes deeper only after reaching a high level of hash table hits.


"AlphaZero vs Stockfish: 25 win for AlphaZero, 25 draw, 0 loss (each program was given 1 minute of thinking time per move, strongest skill level using 64 threads and a hash size of 1GB)"

This is sci-fi. I do not have a 64 core machine but on my pc Stockfish do not sacrifice a Knight for 2 pawns:
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6 4.d3 Bc5 5.Bxc6 dxc6 6.O-O Nd7 7.Nbd2 O-O 8.Qe1 f6 9.Nc4 Rf7 10.a4 Bf8 11.Kh1 Nc5 12.a5 Ne6 13.Ncxe5?
The paper is very interesting. Nevertheless selecting only wins and stripping off all game infos from the pgn might do for non-chess scientists,
but here it is quite useless and remains doubtful.

I hope there is more to come with more details for the games and the setup.
Guenther Simon
http://rwbc-chess.de/

mar
Posts: 1832
Joined: Fri Nov 26, 2010 1:00 pm

Re: Google's AlphaGo team has been working on chess

Post by mar » Wed Dec 06, 2017 11:32 am

While this is indeed incredible, show me how it beats SF dev with good book and syzygy on equal hardware in a 1000 game match.

Alternatively winning next TCEC should do :wink:

Daniel Shawul
Posts: 3437
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Google's AlphaGo team has been working on chess

Post by Daniel Shawul » Wed Dec 06, 2017 12:03 pm

Most of us here suspected that this could happen once Giraffe showed it can beat Stockfish's eval.

Just the fact that the new approch to chess programming worked incredibly well is fantastic even if it didn't beat the best.

Daniel

Post Reply