ChessUSA.com TalkChess.com
Hosted by Your Move Chess & Games
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

CLOP for Noisy Black-Box Parameter Optimization
Post new topic    TalkChess.com Forum Index -> Computer Chess Club: Programming and Technical Discussions Flat
View previous topic :: View next topic  
Author Message
Don Dailey



Joined: 29 Apr 2008
Posts: 5106

PostPost subject: Re: CLOP for Noisy Black-Box Parameter Optimization    Posted: Wed Sep 07, 2011 6:09 pm Reply to topic Reply with quote

Rémi Coulom wrote:
Don wrote:
Rémi Coulom wrote:

Comments and questions are welcome.

Rémi


Larry and I are playing around with CLOP using clop-gui.

In the Max tab, there appears the columns Mean, Max-1, Max-2 and apparently as many columns as there are numbers of parameters to be tuned.

However nothing ever appears in those columns after a day of running. What do those columns mean (no pun.)

I explained earlier in this thread: it is a feature I did not implement. Only max-1 has a meaning: it is the maximum of the quadratic regression, if it exists. It sometimes does not exist. Mean is the weighted mean of all samples. I recommend using parameter values in the "Mean" column for your parameters. They usually perform better according to my artificial experiments.

Quote:
In the paper you mention that this is supposed to converge - so given enough games and time will this eventually stop running? I told Larry that even with a very efficient algorithm it is likely to take millions of games to see values that improve on already well selected values.

It does not converge in a finite amount of games. It converges in the sense that expected simple regret decreases in time, and can be made arbitrarily small with a big enough number of games. Note that I do not have a mathematical proof of this convergence, but I am convinced it does converge if the function to be optimized is continuous.

In the paper I ran experiments on artificial problems. It gives an idea of how close to optimal parameters you get with a given number of games. Around 50% winning rate, one Elo point is about 0.0014. For the Log problem in one dimension, it takes about 10k games to have 0.001 expected simple regret (so, less than one Elo point). For the 5-dimensional problem, it takes a few hundred thousand games to reach 1-Elo expected simple regret (that's 0.2 Elo per parameter).

The magic of CLOP is that it approaches optimal performance faster than it can be measured. Simple regret is smaller than the confidence interval around the estimated win rate.

Rémi


Thanks for explaining that Rémi. I have a script for chess if anyone wants to use it for chess. The script does not have a legal move generator for simplicity and therefore draw and win detection is a bit of a kludge, but it works well.
Back to top
View user's profile Send private message Send e-mail
Display posts from previous:   
Subject Author Date/Time
CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Thu Sep 01, 2011 10:32 am
      Re: CLOP for Noisy Black-Box Parameter Optimization Joona Kiiski Fri Sep 02, 2011 8:52 am
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Sep 02, 2011 9:45 am
                  Re: CLOP for Noisy Black-Box Parameter Optimization Marco Costalba Fri Sep 02, 2011 5:34 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Zach Wegner Fri Sep 02, 2011 4:44 pm
                  Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Sep 02, 2011 7:58 pm
      Re: CLOP for Noisy Black-Box Parameter Optimization Rein Halbersma Fri Sep 02, 2011 10:07 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sat Sep 03, 2011 9:24 am
                  Re: CLOP for Noisy Black-Box Parameter Optimization Rein Halbersma Sat Sep 03, 2011 5:01 pm
                        Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sat Sep 03, 2011 5:28 pm
      Re: CLOP for Noisy Black-Box Parameter Optimization Andres Valverde Sat Sep 03, 2011 10:28 am
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sat Sep 03, 2011 10:40 am
      Re: CLOP for Noisy Black-Box Parameter Optimization Peter Österlund Sat Sep 03, 2011 6:54 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sat Sep 03, 2011 7:34 pm
                  Re: CLOP for Noisy Black-Box Parameter Optimization Ernest Bonnem Sun Sep 04, 2011 4:49 pm
      Re: CLOP for Noisy Black-Box Parameter Optimization Don Dailey Wed Sep 07, 2011 4:57 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Wed Sep 07, 2011 5:47 pm
                  Re: CLOP for Noisy Black-Box Parameter Optimization Don Dailey Wed Sep 07, 2011 6:09 pm
                        Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Wed Sep 07, 2011 6:26 pm
                        Re: CLOP for Noisy Black-Box Parameter Optimization F. Bluemers Wed Sep 07, 2011 6:39 pm
      Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Wed Sep 07, 2011 8:19 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Thu Sep 08, 2011 4:44 pm
                  Re: CLOP for Noisy Black-Box Parameter Optimization Larry Kaufman Fri Sep 09, 2011 3:36 am
                        Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Sep 09, 2011 6:11 am
                              Re: CLOP for Noisy Black-Box Parameter Optimization Larry Kaufman Fri Sep 09, 2011 2:43 pm
                                    Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Sep 09, 2011 7:47 pm
                                          Re: CLOP for Noisy Black-Box Parameter Optimization Marco Costalba Sun Nov 06, 2011 8:24 am
                                                Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Nov 06, 2011 8:30 am
                                          Re: CLOP for Noisy Black-Box Parameter Optimization Marco Costalba Sun Nov 06, 2011 8:50 am
                                                Re: CLOP for Noisy Black-Box Parameter Optimization Peter Österlund Sun Nov 06, 2011 1:11 pm
                                                      Re: CLOP for Noisy Black-Box Parameter Optimization Marco Costalba Sun Nov 06, 2011 4:53 pm
                                                      Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Nov 06, 2011 5:14 pm
                                                            Re: CLOP for Noisy Black-Box Parameter Optimization Peter Österlund Sun Nov 06, 2011 11:11 pm
                                                                  Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Mon Nov 07, 2011 11:18 am
                                                Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Nov 06, 2011 5:05 pm
                                    Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Sep 09, 2011 7:52 pm
                                          Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sat Sep 10, 2011 9:49 am
                                                Re: CLOP for Noisy Black-Box Parameter Optimization Michel Van den Bergh Sun Sep 25, 2011 8:18 am
                                                      Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Sep 25, 2011 8:47 am
                                                            Re: CLOP for Noisy Black-Box Parameter Optimization Michel Van den Bergh Mon Sep 26, 2011 10:55 am
                                                                  Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Mon Sep 26, 2011 1:40 pm
                                                                        Re: CLOP for Noisy Black-Box Parameter Optimization Michel Van den Bergh Tue Sep 27, 2011 7:54 am
                                                                              Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Tue Sep 27, 2011 8:39 am
                                                                                    Re: CLOP for Noisy Black-Box Parameter Optimization Michel Van den Bergh Thu Sep 29, 2011 2:04 pm
                                                                                          Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Thu Sep 29, 2011 3:38 pm
                                                                        Re: CLOP for Noisy Black-Box Parameter Optimization Daniel Shawul Mon Oct 17, 2011 2:03 am
                                                                              Re: CLOP for Noisy Black-Box Parameter Optimization Michel Van den Bergh Mon Oct 17, 2011 7:51 am
                                                                              Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Mon Oct 17, 2011 7:56 am
                                                                                    Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Mon Oct 17, 2011 8:02 am
                                                                                          Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Mon Oct 17, 2011 8:11 am
                                                                                    Re: CLOP for Noisy Black-Box Parameter Optimization Daniel Shawul Mon Oct 17, 2011 9:54 am
                                                                                          Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Mon Oct 17, 2011 11:10 am
      Re: CLOP for Noisy Black-Box Parameter Optimization david nash Wed Oct 05, 2011 12:19 am
      Re: CLOP - a trial for xiangqi engine Edward Yu Fri Oct 07, 2011 11:58 am
            Re: CLOP - a trial for xiangqi engine Rémi Coulom Fri Oct 07, 2011 1:17 pm
                  Re: CLOP - a trial for xiangqi engine Edward Yu Fri Oct 07, 2011 4:01 pm
                        Re: CLOP - a trial for xiangqi engine Rémi Coulom Fri Oct 07, 2011 4:30 pm
                              Re: CLOP - a trial for xiangqi engine Edward Yu Sat Oct 08, 2011 3:59 am
                                    Re: CLOP - a trial for xiangqi engine Rémi Coulom Sat Oct 08, 2011 7:49 am
                                          Re: CLOP - a trial for xiangqi engine Edward Yu Tue Oct 11, 2011 4:04 pm
                                                Re: CLOP - a trial for xiangqi engine Rémi Coulom Tue Oct 11, 2011 4:22 pm
                                                      Re: CLOP - a trial for xiangqi engine Daniel Shawul Tue Oct 11, 2011 4:49 pm
                                                            Re: CLOP - a trial for xiangqi engine Rémi Coulom Tue Oct 11, 2011 5:27 pm
                                                                  Issue.... Michel Van den Bergh Sat Oct 15, 2011 2:51 pm
                                                                        Re: Issue.... Rémi Coulom Sun Oct 16, 2011 6:30 am
                                                                              Re: Issue.... Michel Van den Bergh Sun Oct 16, 2011 8:10 am
                                                                                    Re: Issue.... Rémi Coulom Sun Oct 16, 2011 9:02 am
                                                                                          Re: Issue.... Michel Van den Bergh Sun Oct 16, 2011 3:10 pm
                                                                                          Re: Issue.... Rémi Coulom Sun Oct 16, 2011 7:25 pm
                              Re: CLOP Michel Van den Bergh Sat Oct 08, 2011 6:00 am
                                    Re: CLOP Rémi Coulom Sat Oct 08, 2011 7:53 am
                                          Re: CLOP Rémi Coulom Sat Oct 08, 2011 8:27 am
      Re: CLOP for Noisy Black-Box Parameter Optimization Ferdinand Mosca Tue Oct 11, 2011 8:22 am
            Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Tue Oct 11, 2011 1:38 pm
                  Re: CLOP for Noisy Black-Box Parameter Optimization Ferdinand Mosca Sun Oct 23, 2011 7:39 pm
                        Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Oct 23, 2011 8:55 pm
                              Re: CLOP for Noisy Black-Box Parameter Optimization Ferdinand Mosca Sun Oct 23, 2011 10:47 pm
                                    Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Fri Nov 04, 2011 9:42 am
                                          Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Nov 04, 2011 9:58 am
                                                Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Fri Nov 04, 2011 10:17 am
                                                      Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Nov 04, 2011 10:30 am
                                                            Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Fri Nov 04, 2011 10:44 am
                                                                  Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Nov 04, 2011 10:56 am
                                                                        Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Fri Nov 04, 2011 11:04 am
                                                                              Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Fri Nov 04, 2011 11:56 am
                                                                                    Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Fri Nov 04, 2011 4:16 pm
                                                                                          Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Fri Nov 04, 2011 5:48 pm
      Re: CLOP for Noisy Black-Box Parameter Optimization Sergei S. Markoff Fri Nov 04, 2011 3:02 pm
      Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sat Nov 05, 2011 7:45 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Sat Nov 05, 2011 10:49 pm
            Re: CLOP for Noisy Black-Box Parameter Optimization Ilari Pihlajisto Sat Nov 05, 2011 11:17 pm
                  Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Sun Nov 06, 2011 12:06 am
                        Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Nov 06, 2011 8:00 am
                              Re: CLOP for Noisy Black-Box Parameter Optimization Jonatan Pettersson Sun Nov 06, 2011 10:20 am
                                    Re: CLOP for Noisy Black-Box Parameter Optimization Rémi Coulom Sun Nov 06, 2011 5:10 pm
Post new topic    TalkChess.com Forum Index -> Computer Chess Club: Programming and Technical Discussions

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum




Powered by phpBB © 2001, 2005 phpBB Group
Enhanced with Moby Threads