So, how many of you are working on neural networks for 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.

So, how many of you are working on neural networks for chess?

Poll ended at Mon Mar 04, 2019 6:03 pm

Yep - i am in
0
No votes
No - nothing for me
2
100%
Hm - still waiting for...
0
No votes
 
Total votes: 2

smatovic
Posts: 673
Joined: Wed Mar 10, 2010 9:18 pm
Location: Hamburg, Germany
Full name: Srdja Matovic
Contact:

So, how many of you are working on neural networks for chess?

Post by smatovic » Fri Feb 01, 2019 6:03 pm

It seems to me, that the whole CCC forum is waiting for LC0 to pass Stockfish finally,
so how many of you programmers are working on neural networks for chess?

Myself still ponders on how to combine an AlphaBeta searcher with gpgpu ann eval...

--
Srdja

Robert Pope
Posts: 479
Joined: Sat Mar 25, 2006 7:27 pm

Re: So, how many of you are working on neural networks for chess?

Post by Robert Pope » Fri Feb 01, 2019 10:21 pm

I have to wait until I can get a new computer first. But, hopefully this summer.

AlvaroBegue
Posts: 917
Joined: Tue Mar 09, 2010 2:46 pm
Location: New York
Full name: Álvaro Begué (RuyDos)

Re: So, how many of you are working on neural networks for chess?

Post by AlvaroBegue » Fri Feb 01, 2019 11:22 pm

smatovic wrote:
Fri Feb 01, 2019 6:03 pm
Myself still ponders on how to combine an AlphaBeta searcher with gpgpu ann eval...
I have a promising idea for that, but I don't really have the time to implement it. My job, my wife and my kids are getting in the way of the really important stuff! :)

Daniel Shawul
Posts: 3612
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: So, how many of you are working on neural networks for chess?

Post by Daniel Shawul » Fri Feb 01, 2019 11:30 pm

AlvaroBegue wrote:
Fri Feb 01, 2019 11:22 pm
smatovic wrote:
Fri Feb 01, 2019 6:03 pm
Myself still ponders on how to combine an AlphaBeta searcher with gpgpu ann eval...
I have a promising idea for that, but I don't really have the time to implement it. My job, my wife and my kids are getting in the way of the really important stuff! :)
Watch it you don't who might be surfing the web :)

jorose
Posts: 232
Joined: Thu Jan 22, 2015 2:21 pm
Location: Zurich, Switzerland
Full name: Jonathan Rosenthal

Re: So, how many of you are working on neural networks for chess?

Post by jorose » Fri Feb 01, 2019 11:50 pm

One of the main reasons I am not working on an NN based engine is that I feel I don't have the resources. I think very few people have the resources to seriously train and test a NN based engine.
Reinforcement learning based NN engines are even worse in that regard.

I don't understand how it matters whether Leela or SF is stonger. Depends on the conditions anyways.
-Jonathan

jdart
Posts: 3731
Joined: Fri Mar 10, 2006 4:23 am
Location: http://www.arasanchess.org

Re: So, how many of you are working on neural networks for chess?

Post by jdart » Sat Feb 02, 2019 2:41 am

I think still NN engines even on higher-end commercial graphics cards play blunders at a pretty high rate.

I think there is probably room for a hybrid approach where maybe the NN is suggesting moves to a deep searcher, but I am not aware of anyone pursuing that.

Personally I have a pretty long to-do list for my non-NN engine and I am not planning to drop everything and start building a NN.

--Jon

brianr
Posts: 338
Joined: Thu Mar 09, 2006 2:01 pm

Re: So, how many of you are working on neural networks for chess?

Post by brianr » Sat Feb 02, 2019 3:11 am

jorose wrote:
Fri Feb 01, 2019 11:50 pm
One of the main reasons I am not working on an NN based engine is that I feel I don't have the resources. I think very few people have the resources to seriously train and test a NN based engine.
Reinforcement learning based NN engines are even worse in that regard.

I don't understand how it matters whether Leela or SF is stonger. Depends on the conditions anyways.
My resources include a 2 year-old GTX 1070 GPU. With that I was able to train (supervised learning) a 10x128 Leela net from the "standard CCRL" dataset (link below). It took about a week, and it is competitive with Crafty on between 2 and 4 CPUs, so about 2,900 Elo. It is not as serious as the larger Leela project, but has been quite rewarding learning enough to do it.

http://blog.lczero.org/2018/09/a-standard-dataset.html

Daniel Shawul
Posts: 3612
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: So, how many of you are working on neural networks for chess?

Post by Daniel Shawul » Sat Feb 02, 2019 6:00 am

jdart wrote:
Sat Feb 02, 2019 2:41 am
I think still NN engines even on higher-end commercial graphics cards play blunders at a pretty high rate.

I think there is probably room for a hybrid approach where maybe the NN is suggesting moves to a deep searcher, but I am not aware of anyone pursuing that.

Personally I have a pretty long to-do list for my non-NN engine and I am not planning to drop everything and start building a NN.

--Jon
My NN engine never makes a blunder because I spend 20% of the time doing a mulitpv search calculating scores for root moves,
and then combining that with MCTS scores like: 0.2 * ABscore + 0.8 * MCTSscore. That way the selection of moves at the root is biased
by the alphabeta prior score and that makes it avoid almost all blunders.

Henk
Posts: 5563
Joined: Mon May 27, 2013 8:31 am

Re: So, how many of you are working on neural networks for chess?

Post by Henk » Sat Feb 02, 2019 10:07 am

NN = far too slow = black box = useless

jorose
Posts: 232
Joined: Thu Jan 22, 2015 2:21 pm
Location: Zurich, Switzerland
Full name: Jonathan Rosenthal

Re: So, how many of you are working on neural networks for chess?

Post by jorose » Sat Feb 02, 2019 1:38 pm

Daniel Shawul wrote:
Sat Feb 02, 2019 6:00 am
jdart wrote:
Sat Feb 02, 2019 2:41 am
I think still NN engines even on higher-end commercial graphics cards play blunders at a pretty high rate.

I think there is probably room for a hybrid approach where maybe the NN is suggesting moves to a deep searcher, but I am not aware of anyone pursuing that.

Personally I have a pretty long to-do list for my non-NN engine and I am not planning to drop everything and start building a NN.

--Jon
My NN engine never makes a blunder because I spend 20% of the time doing a mulitpv search calculating scores for root moves,
and then combining that with MCTS scores like: 0.2 * ABscore + 0.8 * MCTSscore. That way the selection of moves at the root is biased
by the alphabeta prior score and that makes it avoid almost all blunders.
How is ScorpioNN coming along?
-Jonathan

Post Reply