Lc0 saturation.

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
stavros
Posts: 136
Joined: Tue Dec 02, 2014 12:29 am

Lc0 saturation.

Post by stavros » Sat Aug 18, 2018 8:24 am

after 37 million games i didnt see any saturation ,maybe that network is better than A0?
can anyone explain this?

frankp
Posts: 216
Joined: Sun Mar 12, 2006 2:11 pm

Re: Lc0 saturation.

Post by frankp » Sat Aug 18, 2018 9:02 am

I certainly cannot comment on the question.
Practically, they would would have to run leela on the same hardware as A0 to know. Although scaling arguments (the leela ratio) have been made on the leela forum to suggest it is not there yet, but also not very far away.
IIRC, David Silver in one of his lectures indicated that A0 was still getting stronger when they stopped the A0 (Go, chess, shogi) training. So it seems A0 could have been, or could be, made even stronger. Their publically stated goal was to demonstrate the potential of their AI approach in these closed systems and that goal had been achieved, so they stopped - apparently.
Last edited by frankp on Sat Aug 18, 2018 9:25 am, edited 1 time in total.

Uri Blass
Posts: 8531
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Lc0 saturation.

Post by Uri Blass » Sat Aug 18, 2018 9:14 am

stavros wrote:
Sat Aug 18, 2018 8:24 am
after 37 million games i didnt see any saturation ,maybe that network is better than A0?
can anyone explain this?
A0 scaled better than stockfish8 based on what I read(lost against stockfish at fast time control only to win at longer time control).
Does lc0 scale better than stockfish8?

If not then it seems that lc0 and A0 do not do the same thing.

jp
Posts: 707
Joined: Mon Apr 23, 2018 5:54 am

Re: Lc0 saturation.

Post by jp » Sat Aug 18, 2018 9:47 am

frankp wrote:
Sat Aug 18, 2018 9:02 am
IIRC, David Silver in one of his lectures indicated that A0 was still getting stronger when they stopped the A0 (Go, chess, shogi) training. So it seems A0 could have been, or could be, made even stronger. Their publically stated goal was to demonstrate the potential of their AI approach in these closed systems and that goal had been achieved, so they stopped - apparently.
No, it was supposedly still getting better for Go, not for chess. Their figure showed it had flatlined for chess after about 12 million games. Or 21 million if you want to be cautious.

stavros
Posts: 136
Joined: Tue Dec 02, 2014 12:29 am

Re: Lc0 saturation.

Post by stavros » Sat Aug 18, 2018 9:52 am

what i mean saturation is lower elo gain per week for example until no further improving

Leo
Posts: 807
Joined: Fri Sep 16, 2016 4:55 pm
Location: USA/Minnesota
Full name: Leo

Re: Lc0 saturation.

Post by Leo » Sat Aug 18, 2018 2:36 pm

I think it is saturated.
Advanced Micro Devices fan.

stavros
Posts: 136
Joined: Tue Dec 02, 2014 12:29 am

Re: Lc0 saturation.

Post by stavros » Sat Aug 18, 2018 9:52 pm

from testserver.lczero.org there is a lot regression, see from 10854 to 10879
http://testserver.lczero.org/matches

whereagles
Posts: 560
Joined: Thu Nov 13, 2014 11:03 am

Re: Lc0 saturation.

Post by whereagles » Mon Aug 20, 2018 12:28 pm

it will eventually saturate the 20x256 NN.. might saturate at a higher ELO with a bigger net, but that's not clear

as of now saturation can be checked for with some statistical time-series test, but just from looking at it, I doubt it reached that point.

JJJ
Posts: 1285
Joined: Sat Apr 19, 2014 11:47 am

Re: Lc0 saturation.

Post by JJJ » Mon Aug 20, 2018 3:32 pm

I heard bigger net was coming anyway. I don't think Lczero will reach the level of the big 3 with this size of net, but it might with the next one. Anyway, in one year, Lczero will be the best engine in the world.

Leo
Posts: 807
Joined: Fri Sep 16, 2016 4:55 pm
Location: USA/Minnesota
Full name: Leo

Re: Lc0 saturation.

Post by Leo » Mon Aug 20, 2018 5:59 pm

JJJ wrote:
Mon Aug 20, 2018 3:32 pm
I heard bigger net was coming anyway. I don't think Lczero will reach the level of the big 3 with this size of net, but it might with the next one. Anyway, in one year, Lczero will be the best engine in the world.
When?
Advanced Micro Devices fan.

Post Reply