[computer-go] anyone applied ideas from Modelling Uncertainty in the Game of Go?
Has anyone applied the ideas in Modelling Uncertainty in the Game of Go by Stern, Graepel, and MacKay?The paper can be found at: http://research.microsoft.com/~dstern/papers/sterngraepelmackay04.pdf It was quite a fascinating paper! - George ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Super-duper computer
This beast goes online in 2011. Better start lobbying now for some Mogo time. http://www.networkworld.com/community/node/32152 ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] CGOS 9x9 seems to be stuck
It looks like it hasn't scheduled any games for the last few hours. David ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] cgos 9x9 is back up, but without anchors.
___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Lockless hash table and other parallel search ideas
RĂ©mi Coulom: [EMAIL PROTECTED]: snip I'll run tests to try to figure out how much strength is lost by parallelization (ie, what is the winning rate of 10,000 sequential playouts vs 1,000 playouts over 10 processors). Hideki ran similar tests against GNU Go, and found 25 Elo loss with 4 CPUs. So 54,193 playouts per second over 16 CPUs will certainly not perform as well as 54,193 sequential playouts. By my recent experiments, 8~9 * (threads - 1) ELO is lost. This matches my earlier result well. Hideki -- [EMAIL PROTECTED] (Kato) ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Super-duper computer
This beast goes online in 2011. Better start lobbying now for some Mogo time. http://www.networkworld.com/community/node/32152 By coincidence I was looking at the Top 500 list yesterday and the top machine already does petaflop (peak) performance [1]. I wonder how many playouts/second Mogo would do on that :-). Darren [1]: http://www.top500.org/blog/2008/06/14/preview_31st_top500_list_world_s_most_powerful_supercomputers_topped_world_s_first_petaflop_s_system -- Darren Cook, Software Researcher/Developer http://dcook.org/mlsn/ (English-Japanese-German-Chinese-Arabic open source dictionary/semantic network) http://dcook.org/work/ (About me and my work) http://darrendev.blogspot.com/ (blog on php, flash, i18n, linux, ...) ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] semeai
Raises hand. Chinese rules version for 9x9 and 13x13 would be quite helpful if that's what you are offering. Different komi would be fine. - Dave Hillis -Original Message- From: Darren Cook [EMAIL PROTECTED] To: computer-go computer-go@computer-go.org Sent: Sat, 6 Sep 2008 10:33 pm Subject: Re: [computer-go] semeai Has anyone tried implementing the ideas in Richard Hunter's Counting Liberties No, but I did make a test suite that included many of the interesting positions (it also included many others of my own creation, both for semeai and tactical search). (And, though almost all the positions were modified enough from those in the book to probably not be covered by copyright, I did get Richard's permission to include them in the test suite.) All positions were modified to fit on a 9x9 board, and IIRC all were set up so that the correct move(s) won the game, while all incorrect moves lost the game. (However that'd be in Japanese rules; they'll need modifying or a different komi for the same to be true in Chinese rules.) If there is any interest I'll dust them off and release the test suite properly. (And would there be interest in the same kind of test suite for 13x13 and 19x19 boards?) Robert Jasiek wrote: - He overlooks some details like exceptional cases and liberty counts for approach defects. (And he does not mention the trivial cases at all.) - It may be suitable to include all LD types in the classification. - For non-trivial types, algorithms need to be developed in addition. - Semeais with kos are more complicated than semeais without kos by a factor roughly between 100 and 1000. Firstly the number of cases explodes (already when there is only one basic ko). Secondly [endgame value] evaluation becomes much more difficult. Thirdly and obviously there are also peculiar ko shapes. Hi Robert, do you have anything published? Or (even better from my point of view) do you have some example positions that demonstrate the exceptional cases, with the correct answer shown? Darren ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] semeai
Darren Cook wrote: do you have anything published? I will do this commercially, i.e., publish when it will be ready and complete. Since currently I am working on projects with a higher priority (among them: joseki books), you need to be very patient, I am afraid. [Unless (email me for $details$)...] Or (even better from my point of view) do you have some example positions that demonstrate the exceptional cases, with the correct answer shown? These hints I shall reveal: - Hunter does not discuss trivial cases like, e.g., big eye with internal string where the opponent's major string is on atari. - Hunter overlooks some exceptional cases about smallest possible numbers of liberties for one side like, e.g., 0 or 1 external liberties on one side. Such cases do not occur for all major types but only for some major types. (In semeais with kos, there are more such exceptional cases.) -- robert jasiek ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/