Re: [Computer-go] UEC Cup will be held tomorrow
Hi, After final 5 round, VisionGo won. 1st VisionGo 5-0 2nd RankaGo 3-2 3rd KataGo3-2 4th DaPangGo 3-2 5th ChaoRanGo 1-5 6th Kohada0-6 Ran Kat Vis Cha Koh Dap Wins Rank RankaGo1 0 1 1 0 32 KataGo 0 0 1 1 1 33 VisionGo 1 1 1 1 1 51 ChaoRanGo 0 0 0 1 0 15 Kohada 0 0 0 0 0 06 DaPangGo 1 0 0 1 1 34 Rn CS Ray mar Aya Qin Wins Rank Rn 1 1 0 1 1 47 CrazyStone 0 1 1 1 1 48 Ray0 0 0 1 1 2 11 marble 1 0 1 0 1 39 Aya0 0 0 1 1 2 10 QinoaIgo 0 0 0 0 0 0 12 Thanks, Hiroshi Yamashita On 2021/03/19 21:27, Hiroshi Yamashita wrote: Hi, 12th UEC Cup will be held tomorrow. First day, there are 7 swiss round for preliminary. Second day, Top 6 programs will play final 5 rounds. From 7th to 12th will play 5 rounds in another league. UEC Cup http://entcog.c.ooco.jp/entcog/new_uec/en/ Game live on youtube https://www.youtube.com/channel/UCt5CU4KCMtsQQmG2v3q51Qw/ Game live on web (updated every 30 seconds) http://www.yss-aya.com/uec2021/ A commentary session will be held on the 21st. (in Japanese). http://entcog.c.ooco.jp/entcog/new_uec/watch.html There will be 19 participants. (Maru is canceled.) http://entcog.c.ooco.jp/entcog/new_uec/en/list.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] UEC Cup will be held tomorrow
Hi, 12th UEC Cup will be held tomorrow. First day, there are 7 swiss round for preliminary. Second day, Top 6 programs will play final 5 rounds. From 7th to 12th will play 5 rounds in another league. UEC Cup http://entcog.c.ooco.jp/entcog/new_uec/en/ Game live on youtube https://www.youtube.com/channel/UCt5CU4KCMtsQQmG2v3q51Qw/ Game live on web (updated every 30 seconds) http://www.yss-aya.com/uec2021/ A commentary session will be held on the 21st. (in Japanese). http://entcog.c.ooco.jp/entcog/new_uec/watch.html There will be 19 participants. (Maru is canceled.) http://entcog.c.ooco.jp/entcog/new_uec/en/list.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] 12 UEC Cup in March
Hi, UEC Cup will be held in 20-21th March. It is an online event this year. Application deadline is 14th February. You can use a software with unique improvements to KataGo or LeelaZero. Current participants are http://entcog.c.ooco.jp/entcog/new_uec/en/list.html Thanks, Hiroshi Yamashita Rémi Coulom Thu, 21 Jan 2021 15:49:11 -0800 Hi, I had not noticed, but there will be an online UEC Cup in March: http://entcog.c.ooco.jp/entcog/new_uec/ It seems that the English page there has the wrong date: http://entcog.c.ooco.jp/entcog/new_uec/en/outline.html If I understand correctly it will take place on March 20th-21st. I will register. Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Leela Zero is playing self-atari in seki
Hi Remi, Leela was winning but lost the game by playing self-atari in a seki: It looks like this only happens when using "kgs-genmove_cleanup". I use this command to remove dead stones. "genmove" does not play this kind of moves. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGOS source on github
Hi, The most noticeable case of this is with Mi Yuting's flying dagger joseki. I'm not familiar with this. I found Hirofumi Ohashi 6d pro's explanation half year ago in HCCL ML. The following is a quote. - https://gokifu.net/t2.php?s=3591591539793593 It seems that it is called a flying dagger joseki in China. This shape, direct 33 to lower tsuke (black 9th move B6) is researched jointly with humans and AI, but still inconclusive. After kiri (black 15th move E4), mainstream is white A, but depending on the version of KataGo, white B may be recommended. By the way, KataGo I'm using now is 1.3.5, which is just a short time ago. This kind of joseki is not good for Zero type. Ladder and capturing race are intricately combined. In AlphaGo(both version of AlphaGoZero and Master) published self-matches, this joseki is rare. - I found this joseki in kata1_b40s575v100 (black) vs LZ_286_e6e2_p400 (white). http://www.yss-aya.com/cgos/viewer.cgi?19x19/SGF/2021/01/22/733340.sgf Mi Yuting wiki has this joseki. https://zh.wikipedia.org/wiki/%E8%8A%88%E6%98%B1%E5%BB%B7 KataGo has special option. https://github.com/lightvector/KataGo/blob/4a79cde56e81209ce4e2fd231b0f2cbee3a8354b/cpp/neuralnet/nneval.cpp#L499 a very large sampling of positions from a wide range of human professional games, from say, move 20, and have bots play starting from these sampled positions, in pairs once with each color. This sounds interesting. I will think about another CGOS that handle this. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGOS source on github
Hi, This is original BayesElo. I updated manually. This is latest. http://www.yss-aya.com/cgos/19x19/bayes.html CrazyStone-18.044065 CrazyStone-81b-TiV 4032 Zen-15.7-3c1g 3999 CrazyStone-57-TiV 3618 This renames CrazyStone-57-TiV to CrazyStone-18.04. http://www.yss-aya.com/cgos/19x19/bayes_20210121_rename_CrazyStone-57-TiV_to_CrazyStone-18.04.html CrazyStone-81b-TiV 4051 Zen-15.7-3c1g 3968 CrazyStone-18.043778 Would you like to continue to rename CrazyStone-57-TiV? The drift looks fixed, but a little for other programs. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGOS source on github
Hi, The Elo statistical model is wrong when different kind of programs play against I have a similar experience. I calculated Japanese Shogi women pro rating before. The strongest woman, Ichiyo Shimizu, her rating is 1578 Elo. Her winrate against men pros is 18%(163 games), and against women pro is 65%(523 games). Her rating without women pros game is 1286 Elo. There is 292(=1578 - 1286) Elo difference. It is because usually women pros play with women pros. Women pros vs men pros are rare. I think similar thing happens on CGOS. There are three eras, Zen, LeelaZero and KataGo. Number of Zen vs LeelaZero games are a little. CrazyStone-18.04 rate maybe depend on Zen-15.7-3c1g. http://www.yss-aya.com/cgos/19x19/cross/CrazyStone-18.04.html Zen's absence is maybe a reason of this drift. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGOS source on github
Hi,
I have published current CGOS source on github.
https://github.com/yssaya/CGOS
There are some changes. Like
1. Recent 300 games on cross-table page.
2. WGo viewer
3. 7.0 komi and handling draw for rating calculation.
4. Shorter pgn file for BayesElo (cgosBayes).
5. Forbid number only account.
6. Bug fixed to send info all 'catch {puts $soc "info $msg"}'
7. badusers.txt for not removing dead stones or too many timeout.
Thanks,
Hiroshi Yamashita
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGOS rating has drifted a lot
Hi Remi, This has latest 19x19.pgn and some tools to make bayes.html. http://www.yss-aya.com/20210112cgos_bin.tar.bz2 I got these tools from Joshua Shriver. Thanks, Hiroshi Yamashita Hi Hiroshi, I tried to compile my old code, but it turns out to be a bit too complicated because it was using tensorflow. I am not using tensorflow any more, and don't want to try to reinstall and recompile it. But I would like to try to take a look at the data. Can I download it anywhere? Can you send the PGN of the 3 lists to me? Thanks, Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGOS rating has drifted a lot
Hi, Deleting losing on time games, and deleting losing on time games and same program matches are follows. 19x19 BayesElo (deleting losing on time games) http://www.yss-aya.com/cgos/19x19/bayes_20210109_del_timeout.html 19x19 BayesElo (deleting losing on time games and same program matches) http://www.yss-aya.com/cgos/19x19/bayes_20210109_del_timeout_same_program_match.html Top rating and ELFv2 (delete on time) black_hole_v75245 LZ_05db_ELFv2_p800 4005 Top rating and ELFv2 (delete on time and same program matches) black_hole_v64958 LZ_05db_ELFv2_p800 3855 Top rating and ELFv2 (original) Katago_40b_s37_4V4762 LZ_05db_ELFv2_p800 3686 http://www.yss-aya.com/cgos/19x19/bayes.html I had thought top rating of "delete on time and same program matches" was less than original, but it is not. The matches between these names are deleted. For example, if black name has "LZ" and white name has "ELF", this game is deleted. LZ | lz | Leela | ELF kata | Kata | KG | k40b | k160 | Ka Zen CrazyStone | CS Rnz | rn | Rn Aya myCtest AQ Dirichlet CGI DCNN- Odin 731369 ... number of games 721122 ... number of games after deleting timeout 592249 ... number of games after deleting timeout and same program matches Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGOS rating has drifted a lot
Hi Remi, I agree with you losing on time against weak program is bad effect for ratings. And I think same program matches with different playouts, like 400 playouts vs 800 playouts are worse. There are too many Leela Zero vs Leela Zero, and recently KataGo vs KataGo. Maybe for human, current top program rating is not so high. I will try to calculate BayesElo without losing on time and same program matches before very long. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGF Open 2020 result
Hi, CGF Open was held in 2020-10-10 and 11. Each players stayed at home. All games are played on NNGS server. In 9x9, 7 players and GNU Go as a guest are played double round robin. And Ray won. Ray is trained by 3 million games. Its NN is 256 filters and 19 resnet(SENet) blocks. Trained games are generated in a similar way to KataGo. It has ownership output and three(win, draw, loss) separate value output. 9x9 result, October 10th, 10 minutes sudden death, Chinese rule, komi 7.0 Ray Rn BSK Koha Aya GNU Pyaq Kats Wins Rank Ray =1 11 11 01 11 11 11 12.51 Rn 0=01 =1 11 11 11 11 11 2 BSK00 01== 01 11 11 119 3 Kohada 00 0= ==1= 11 11 119 4 Aya01 00 01 =011 11 107.55 GNU Go 00 00 00 00 0011 114 6 PyaqArgo 00 00 00 00 00 00112 7 Katsunari 00 00 00 00 10 00 00 1 8 * "=1" means draw for black, win for white. "=" is 0.5 win. In 19x19, 7 players and GNU Go as a guest are played swiss 6 rounds. And Rn won. Kohada is a new program. 19x19 result 2020-10-11, 30 minutes sudden death, Japanese rule, komi 6.5, Online Rank Program Name 1 2 3 4 5 6 Pt SOS SB MD 1 Rn 8+ 3+ 2+ 5+ 6+ 4+ 6.0 17.0 17.0 12.0 2 Kohada 5+ 4+ 1- 3+ 7+ 6+ 5.0 19.0 13.0 8.0 3 BSK 6+ 1- 4+ 2- 8+ 7+ 4.0 17.0 6.0 3.0 4 Ray 7+ 2- 3- 8+ 5+ 1- 3.0 19.0 4.0 1.0 5 Aya 2- 7+ 6+ 1- 4- 8+ 3.0 17.0 3.0 1.0 6 GNU Go 3- 8+ 5- 7+ 1- 2- 2.0 19.0 1.0 0.0 7 Katsunari4- 5- 8+ 6- 2- 3- 1.0 17.0 0.0 0.0 8 PyaqArgo 1- 6- 7- 4- 3- 5- 0.0 19.0 0.0 0.0 SGF http://hp.vector.co.jp/authors/VA012620/cgf2020/cgf2020.zip 2020 CGF Open result(in Japanese) http://hp.vector.co.jp/authors/VA012620/cgf2020/result2020.html CGF Open(in Japanese) http://hp.vector.co.jp/authors/VA012620/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] 2020 World Artificial Intelligence Go Competition
Hi, After preliminary swiss 7 rounds, top 8 programs will play quarterfinal. A bit surprising, LeelaZero were qualified, but KataGo could not. It seems there are many timeout games though. Wins SB Rank Myrtyle Go 6 30 1 ChoRanGo 6 29 2 Golaxy 5 28 3 Tian Go 5 26 4 YiLeGo 4 30 5 DolBaram 4 29 6 LeelaZero4 29 7 izisgo 4 25 8 KataGo 4 25 9 Clear Stone 4 19 10 Yixiaotian 3 28 11 AiqiGo 3 25 12 WeiHuGo 3 25 13 FlyingTiger 3 23 14 Handtalk 3 23 15 DaPangGo 3 19 16 BaduGI 3 18 17 XiaofeiGo2 19 18 Go Sprit 1 21 19 BSK 0 19 20 Result (BSK author) https://twitter.com/oubeika11/status/1310903786483802115 Official? result and games (Chinese) http://game.egcode.cn/#/match/7 2020 World Artificial Intelligence Go Competition https://senseis.xmp.net/?2020WorldArtificialIntelligenceGoCompetition 2020 World Artificial Intelligence Go Competition Rules & Information https://gameofgo.app/general/world-go-artificial-intelligence-tournament-2020 Introduce of this competition (Chinese) https://sports.sina.com.cn/go/2020-09-08/doc-iivhvpwy5522546.shtml Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] draw rate against even opponent in CGOS 9x9
Hi, This graph is draw rate and black winrate against even opponent in CGOS 9x9. http://www.yss-aya.com/20200704cgos9x9.png Blue is black winrate. Red is draw rate. The higher rating, the higher draw rate. If it keeps going this way, all the games will be draw in 4500 ELO. In black winrate, it seems the higher rating, the less than 50%. There are from 11 games from 2020-01-01 to 2020-06-30. For example, draw-rate between programs in their 3300s are 0.444. ELO games b-win draw w-win black winrate draw-rate 1800: 32 =( 18/ 0/ 14),0.563, 0.000 1900:0 =( 0/ 0/ 0),0.000, 0.000 2000: 62 =( 28/ 7/ 27),0.508, 0.113 2100: 296 =(132/ 23/141),0.485, 0.078 2200:0 =( 0/ 0/ 0),0.000, 0.000 2300: 33 =( 14/ 2/ 17),0.455, 0.061 2400: 68 =( 30/ 2/ 36),0.456, 0.029 2500: 122 =( 54/ 10/ 58),0.484, 0.082 2600: 280 =( 98/ 45/137),0.430, 0.161 2700: 51 =( 23/ 9/ 19),0.539, 0.176 2800: 582 =(245/153/184),0.552, 0.263 2900: 1458 =(570/433/455),0.539, 0.297 3000: 554 =(153/196/205),0.453, 0.354 3100: 450 =(116/157/177),0.432, 0.349 3200: 131 =( 34/ 52/ 45),0.458, 0.397 3300: 124 =( 22/ 55/ 47),0.399, 0.444 3400:0 =( 0/ 0/ 0),0.000, 0.000 * GNU GO is 1800 ELO. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGOS max rating graph
Hi, I made CGOS max rating graph. http://www.yss-aya.com/20200703cgos_e.png Red is 19x19, Blue is 9x9. In 2006, Monte-Carlo was introduced, and progressing slowly up to 2012. In 2016, Deep Learning was introduced, and big jump in 19x19. But it looks like to improve 9x9 is very hard. GNU Go is 1800 ELO. In 9x9, komi is 7.5 up to 2012, 7.0 from 2016. 19x19 time limit is 20 minutes up to 2012, 15 minutes from 2016. I use BayesELO if it has. CGOS had stopped from 2012 to 2015. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGOS has two anchors
Hi, LZ_05db_ELFv2_p800 has been fixed at 3670(from 3102) to match BayesElo. This is an adjustment with a difference of 568 Elo between BayesElo and standing page Elo on 19x19. http://www.yss-aya.com/cgos/19x19/standings.html Some bots are also fixed at BayesElo temporarily. If I found old bots which have BayesElo, I will fix them at BayesElo temporarily. On the standing page, GNU Go is fixed at 1800. Now that there are over 4000 Elo bots, many higher scores are being compressed. For example, if ELFv2 is 3600 and this is only one bot at the top, and if you put 4000 KataGo there, ELFv2 drops to around 3550 and KataGo stabilizes at 3950. I think we're repeating this process. LZ_05db_ELFv2_p800 uses LeelaZero 0.15. name LZ_05db_ELFv2_p800 invoke ../leelaz -p800 --noponder -r1 -g -t2 -q -d -w ../networks/05dbca157002b9fd618145d22803beae0f7e4015af48ac50f246b9316e315544 Gnugo-3.7.10-a1 uses ver 3.7.10 and 32 bit binary. 64 bit GNU Go has some problems. name Gnugo-3.7.10-a1 invoke ./gnugo3710_32bit --mode gtp --capture-all-dead --chinese-rules --min-level 10 --max-level 10 --positional-superko Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] rn-6.3.0 released
Hi, Kensuke Matsuzaki released rn-6.3.0. It is one of the strongest 9x9 engine. It is "LeelaZero + 9x9 + heuristic features + adjustable komi + KataGo like learning". rn-6.3.0 https://github.com/zakki/leela-zero/releases/tag/rn-6.3.0 Rn.6.3 https://twitter.com/k_matsuzaki/status/1260908554359173120 Author says v995 is a latest model, but v945 is stronger. And v995 tends to play (4,4) on initial position. Thanks, Hiroshi Yamashita This is quote from README.md in zip file. --- # '9x9-endstate' branch * For 9x9 game. * Ladder detection (by https://github.com/yssaya/leela-zero-ladder) * Various komi (by https://github.com/ihavnoid/leela-zero) * Additional input features. --- # 'Endstate' branch This is a fork of Leela Zero with the 'endstate' head. The 'stock' Leela Zero uses the value and policy nets, while this also predicts how the game ends. To do so, there are some changes: * Additional 'endstate' head : The 'endstate' is how the game ended - that information is also stored on the training data. * Acceleration mode : To predict the endstate, we can't just resign when we find the game hopeless - we have to play it to the end. Hence, once we hit the resignation threshold, we reduce the playouts to 1 instead of resigning. * Using the 'endstate' information as the auxillary policy - see Network.cpp for details on how it uses the auxilary policy The main goal of this branch is to play reasonable handicap games (and to some extent, play games with komi) --- ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] GLOBIS-AQZ has been released
Hi, Yu Yamaguchi has released GLOBIS-AQZ as v4.0. GLOBIS-AQZ https://github.com/ymgaq/AQ v4.0.0 https://github.com/ymgaq/AQ/releases Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Lee Sedol vs HanDol
Hi, Lee Sedol won first game against HanDol with two handicaps. White K7 escape does not work by Black J6 (Geta). This is maybe easy for human, but not for programs. White can chase Black L11 from White L12. But this is not ladder. And White also can play N10. So understanding White K8 death needs may search plies. My old classic program Aya also can not understand this K8 death. a b c d e f g h j k l m n o p q r s t 19 . . . . . . . . . . . . . . . . . . . 19 18 . . . . . . . . . . . . . . . . . . . 18 17 . . . . . . . . . . . X X X X . O X . 17 16 . . . X . . . . . + O X O . O X O X . 16 15 . . . . . . . . . . X O . X X X X O X 15 14 . . X . . . . . . . . O . O O O O O . 14 13 . . . . . . . . . . . . . . . . . O . 13 12 . . . . . . . . . . . . . . . X . . . 12 11 . . . . . . . . . . X O . . O . . . . 11 10 . . . + . . . . O O X O . X X X . . . 10 9 . . . . . . . . X X O X X . . . . . . 9 8 . . . . . . . .(X)O O O X O X . O . . 8 7 . . . . . . . . . . X X O . . X . . . 7 6 . . . . . . . . . . . . . O . . . O . 6 5 . . . X . . . . . . X . . . . X . . . 5 4 . . O X . . . . O O O O . O O X O . . 4 3 . . O X O O . X X X O X . . X O O . . 3 2 . . O O X . . X O O X X . . X . . . . 2 1 . . . . . . . . . . . . . . . . . . . 1 a b c d e f g h j k l m n o p q r s t (;GM[1]SZ[19] PB[Lee Sedol] PW[HanDol] RE[B+R]HA[2]KM[7.5]RU[Japanese]DT[2019-12-18] PC[the Seoul headquarters of Bodyfriend] EV[Lee Sedol retirement match] AB[pd][dp] ;W[qp];B[dd];W[np];B[lq];W[fq];B[pp];W[pq];B[oq];W[op];B[po] ;W[qq];B[or];W[rn];B[iq];W[cq];B[dq];W[cp];B[do];W[dr];B[er] ;W[cr];B[pm];W[eq];B[cf];W[qf];B[nc];W[nn];B[pj];W[nl];B[qe] ;W[pf];B[ph];W[nf];B[ld];W[le];B[ke];W[lf];B[nj];W[oi];B[oj] ;W[md];B[mc];W[kd];B[lc];W[ql];B[ol];W[od];B[oc];W[qc];B[oe] ;W[re];B[rd];W[qd];B[pe];W[rg];B[se];W[rf];B[rc];W[of];B[ne] ;W[jp];B[kr];W[lj];B[mk];W[kl];B[lk];W[kk];B[kj];W[li];B[lm] ;W[lp];B[km];W[ip];B[hq];W[ll];B[ml];W[mm];B[ik];W[jj];B[ki] ;W[ij];B[ko];W[kp];B[jk];W[jr];B[jq];W[kq];B[lr];W[ir];B[hr] ;W[jl];B[il]) Go master Lee Sedol takes on another AI machine before he retires https://pulsenews.co.kr/view.php?year=2019=1014122 Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] UEC 11th, GOLAXY won
Hi, GOLAXY won. Rank 1 Golaxy 5-0 2 GLOBIS-AQZ 3-2 3 BaduGI 3-2 4 Ray 3-2 5 Natsukaze 1-4 6 Maru0-5 * 2,3,4 was decided by preliminary rank http://entcog.c.ooco.jp/entcog/new_uec/en/ (;GM[1]SZ[19] PB[GLOBIS-AQZ] PW[GOLAXY] DT[2019-12-15]RE[W+100.5]KM[6.5]TM[1800]RU[Japanese] PC[AWS Tokyo Loft, Japan]EV[UEC 11th]GN[Final 5R] ;B[dd];W[pc];B[pp];W[cq];B[qe];W[nd];B[qn];W[nq];B[lc];W[qg] ;B[pg];W[ph];B[pf];W[rd];B[me];W[oh];B[qh];W[qi];B[rh];W[pj] ;B[qk];W[ri];B[rg];W[re];B[ng];W[mi];B[lh];W[li];B[kh];W[mc] ;B[md];W[mb];B[lb];W[mh];B[mg];W[ld];B[ne];W[ki];B[ih];W[qq] ;B[pq];W[pr];B[qp];W[rr];B[rq];W[qr];B[fc];W[cf];B[ji];W[bd] ;B[ol];W[ef];B[cc];W[ci];B[eh];W[de];B[fe];W[ff];B[dg];W[cd] ;B[dc];W[bh];B[od];W[ob];B[ee];W[df];B[mq];W[mp];B[nr];W[mr] ;B[lq];W[or];B[np];W[oq];B[jk];W[jg];B[jh];W[pl];B[pk];W[ok] ;B[rk];W[nk];B[om];W[kl];B[oc];W[nb];B[eq];W[ep];B[ml];W[rf] ;B[qf];W[mk];B[fp];W[dp];B[gq];W[lr];B[gf];W[gg];B[fg];W[hg] ;B[hf];W[kc];B[kb];W[jb];B[kd];W[la];B[jc];W[ka];B[le];W[ej] ;B[cg];W[bg];B[dj];W[di];B[en];W[fo];B[eo];W[co];B[go];W[fi] ;B[ll];W[kk];B[ln];W[jl];B[cn];W[bn];B[ik];W[kn];B[lp];W[ig] ;B[bm];W[cm];B[dn];W[bl];B[in];W[qj];B[gh];W[fh];B[eg];W[jq] ;B[mo];W[bc];B[oi];W[lk];B[dm];W[am];B[hh];W[kg];B[bf];W[ce] ;B[jf];W[kf];B[lg];W[if];B[kj];W[lj];B[qc];W[qd];B[pd];W[pb] ;B[rc];W[rb];B[sb];W[qb];B[ie];W[je];B[sc];W[sd];B[sf];W[sa] ;B[lf];W[he];B[ke];W[ge];B[nj];W[oj];B[ni];W[nh];B[og];W[gb] ;B[gc];W[hb];B[hc];W[ib];B[id];W[dl];B[el];W[ek];B[rj];W[fl] ;B[fm];W[gl];B[gm];W[fr];B[er];W[fq];B[gp];W[ir];B[gr];W[sj] ;B[sk];W[si];B[sl];W[op];B[oo];W[nn];B[no];W[nl];B[nm];W[pm] ;B[on];W[rm];B[sm];W[rn];B[sn];W[ro];B[rp];W[so];B[pn];W[ql] ;B[qm];W[rl];B[qo];W[dr];B[fs];W[fb];B[eb];W[cb];B[db];W[ca] ;B[da];W[bb];B[fd];W[ic];B[hd];W[jf];B[jd];W[gj];B[gf];W[hf] ;B[gd];W[gf];B[gi];W[hj];B[ei];W[fj];B[dk];W[cl];B[hk];W[em] ;B[gk];W[el];B[fn];W[lm];B[mm];W[ko];B[km];W[jm];B[lm];W[ip] ;B[jn];W[hn];B[hm];W[im];B[ho];W[io];B[gn];W[jo];B[hn];W[sh] ;B[sg];W[se];B[qg];W[sq];B[sp];W[sr];B[kr];W[kq];B[ls];W[ns] ;B[ms];W[ks];B[ls];W[kp];B[js];W[hr];B[hq];W[lo];B[mp];W[iq] ;B[hs];W[is];B[hp];W[dq];B[gs];W[nc];B[ds];W[cs];B[es];W[oe] ;B[pe];W[jj];B[ij];W[do];B[kj];W[fa];B[ea];W[ec];B[ed];W[jj] ;B[hi];W[kj];B[hl];W[fk];B[il]) Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] UEC 11th first day result
Hi, UEC 11th held in Amazon Loft, Meguro, Tokyo, Japan. 18 program played preliminary swiss 7R yesterday. Top 6 programs will play today's final league. Rank 1 Golaxy China 7-0 2 GLOBIS-AQZ Japan 6-1 3 BaduGI Korea 6-1 4 Natsukaze Japan 5-2 5 RayJapan 5-2 6 Maru Japan 4-3 7 Go Genius China 3.5-3.5 8 nlpJapan 3.5-3.5 9 Akira Japan 3-4 10 BSKJapan 3-4 11 MayoiGoJapan 3-4 12 Rn Japan 3-4 https://twitter.com/UgenNihonkiin/status/1205791218329735168 http://entcog.c.ooco.jp/entcog/new_uec/en/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] The 11th UEC Cup has been announced
Hi, This is reminder. UEC Cup application deadline is November 25. http://entcog.c.ooco.jp/entcog/new_uec/en/ December 14 & 15 Application deadline is November 25. This time, the venue is a bit interesting place. Amazon AWS Loft Tokyo. https://aws.amazon.com/jp/start-ups/loft/tokyo/ (in Japanese) AWS Loft needs personal information for entry. So spectators for commentary event will have to apply in advance. This will start from November 25. And this prize is also interesting. AWS will offer AWS cloud coupon for this award. AWS Award (Encouragement award) We give this award to excellent program with promising future developed under relatively poor development environment. http://entcog.c.ooco.jp/entcog/new_uec/en/prizes.html Thanks, Hiroshi Yamashita Rémi Coulom Sat, 26 Oct 2019 12:09:10 -0700 http://entcog.c.ooco.jp/entcog/new_uec/en/ December 14 & 15 Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] The 11th UEC Cup has been announced
Hi, http://entcog.c.ooco.jp/entcog/new_uec/en/ December 14 & 15 Application deadline is November 25. This time, the venue is a bit interesting place. Amazon AWS Loft Tokyo. https://aws.amazon.com/jp/start-ups/loft/tokyo/ (in Japanese) AWS Loft needs personal information for entry. So spectators for commentary event will have to apply in advance. This will start from November 25. And this prize is also interesting. AWS will offer AWS cloud coupon for this award. AWS Award (Encouragement award) We give this award to excellent program with promising future developed under relatively poor development environment. http://entcog.c.ooco.jp/entcog/new_uec/en/prizes.html Thanks, Hiroshi Yamashita Rémi Coulom Sat, 26 Oct 2019 12:09:10 -0700 http://entcog.c.ooco.jp/entcog/new_uec/en/ December 14 & 15 Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] ML web site was deleted?
Hi, I got "Not found" err. http://computer-go.org/mailman/listinfo/computer-go I can not see past mails too. http://computer-go.org/pipermail/computer-go/ I can see here. http://computer-go.org/ Another past mails are available. https://www.mail-archive.com/computer-go@computer-go.org/maillist.html https://groups.google.com/forum/#!forum/computer-go-archive I haven't received any emails since March. There are some problems. Re: [Computer-go] List problem ... https://www.mail-archive.com/computer-go@computer-go.org/msg17492.html Erik reported in Feb 17 2019, Remi also reported in Mar 24 2019. https://www.mail-archive.com/computer-go@computer-go.org/msg17488.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGF Open 2019 result
Hi, CGF Open2019 was held in The University of Electro-Communications, Chofu, Tokyo, Japan, on 13th-14th July. 8 programs played round-robin in 9x9. Top 3 were 1st Ray14-0 2nd Natsukaze 11-3 3rd AIZE 10-4 10 programs(GNU Go as a guest) played swiss 6 rounds. Top 3 were 1st Raynz 6-0 2nd AIZE 5-1 3rd Natsukaze 4-2 Ray author is Yuki Kobayashi. Ray uses AlphaZero style interesting network on 9x9. Raynz and AIZE author is TRIPLEIZE. Natsukaze author is Tomoyuki Kaneko and Hiroshi Yamashita. 2019-07-13 The University of Electro-Communications, Chofu, Tokyo, Japan West 9, 3F AV hole. round-robin, Chinese, Komi 7.0, 10 minutes Nat nlp Rn Arg BSK Ray Rnz AIZ Wins Rank Natsukaze 11 11 11 11 00 11 01 112 nlp 00 11 01 11 00 00 00 55 Rn 00 00 01 01 00 00 00 28 PyaqArgo00 01 01 01 00 01 00 46 BSK 00 00 01 01 00 10 00 37 Ray 11 11 11 11 11 11 11 141 Raynz 00 11 11 01 10 00 01 74 AIZE01 11 11 11 11 00 01 103 "01" is lose for black, win for white. 2019-07-14 Swiss 6 rounds. Japanese rule. Komi 6.5. 30 minutes. nngs +-+++++++--+--+-++ | Player | 1R | 2R | 3R | 4R | 5R | 6R | Wins |SL|SB/MD|Rank| +-+++++++--+--+-++ | 1.Natsukaze |GNU |BSK |AIZE|nlp |Rayn|Rn | 4-2-0|19| | 3| | | 1| 1| 0| 1| 0| 1| 4.0| | || | 2.Rn|AIZE|GNU |Pyaq|Ray |BSK |Nats| 3-3-0|16| 3.0| 5| | | 0| 1| 1| 0| 1| 0| 3.0| | || | 3.Ray |Rayn|Pyaq|nlp |Rn |AIZE|BSK | 4-2-0|18| | 4| | | 0| 1| 1| 1| 0| 1| 4.0| | || | 4.Maru |BSK |AIZE|Rayn|GNU |nlp |Argo| 3-3-0|16| 3.0| 5| | | 1| 0| 0| 1| 0| 1| 3.0| | || | 5.nlp |Pyaq|Rynz|Ray |Nats|Maru|AIZE| 2-4-0|22| | 7| | | 1| 0| 0| 0| 1| 0| 2.0| | || | 6.PyaqArgo |nlp |Ray |Rn |BSK |GNU |Maru| 0-6-0| | | 10| | | 0| 0| 0| 0| 0| 0| 0.0| | || | 7.BSK |Maru|Nats|GNU |Pyaq|Rn |Ray | 2-4-0|15| | 8| | | 0| 0| 1| 1| 0| 0| 2.0| | || | 8.Raynz |Ray |nlp |Maru|AIZE|Nats|GNU | 6-0-0| | | 1| | | 1| 1| 1| 1| 1| 1| 6.0| | || | 9.AIZE |Rn |Maru|Nats|Rayn|Ray |nlp | 5-1-0| | | 2| | | 1| 1| 1| 0| 1| 1| 5.0| | || |10.GNU Go|Nats|Rn |BSK |Maru|Pyaq|Rayn| 1-5-0|18| | 9| | | 0| 0| 0| 0| 1| 0| 1.0| | || +-+++++++++-++ Note: Round 6 matching was made by hand, because software matching failed. SGF http://hp.vector.co.jp/authors/VA012620/cgf2019/cgf2019.zip CGF Open (in Japanese) http://hp.vector.co.jp/authors/VA012620/ Thanks Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] New UEC Cup will be held on December 2019
Hi, New UEC Cup(tentative name) will be held on 7th-8th or 14th-15th December 2019 in Japan. Next AI Ryusei will not be held. Rule will be same as AI Ryusei. Details are still undecided. I will post here as soon as the details are decided. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Leela Zero for Shogi, AobaZero
Hi, I have a GTX 1080Ti that can do a pretty good job as well :) Yes, using your own GPU is best. But recently I have found Colab and it is so easy. Maybe on this ML, there is many people who don't have GPU. Colab is nice service for them. . Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Leela Zero for Shogi, AobaZero
Hi, We have started Leela Zero for Shogi, AobaZero. http://www.yss-aya.com/aobazero/index_e.html https://github.com/kobanium/aobazero/blob/release/README_en.md Like LZ, it is user-participated, and free open-source AlphaZero project. Anyone can contribute using Google Colab. Run AobaZero on a Tesla T4 for free(Google Colaboratory) http://www.yss-aya.com/aobazero/colab_e.html We use LZ 0.17 OpenCL code for client engine. Server is original. I would like to thank GCP and Leela Zero team for source codes and many issues(include LCZero). They were really useful. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AlphaZero paper difference between 2017 and 2018
Hi Ichikawa san, Thank you for nice explanation. I think your guess is maybe right. And 2018 nature paper might have no mistake. I had checked carefully both Figure 1. 1. 2017 reaches AlphaGo Lee in 170,000 step. 2018 reaches in 80,000 step. 2. 2017 and 2018 reach "AlphaGo Zero(20 block)" in similar steps. 3. Final strength is similar. So I had thought "If you use 7 times games record, initial learning speed is fast, but final strength is similar.". So maybe they want to say "21 million Training Games is enough." But it is wrong. In Go, if you use all positions from a game, it makes overfitting? And learning will fail? Without symmery-augmented, Go can use only 20 positions from a game. Chess and Shogi is ok. It looks like domain dependent... Thanks, Hiroshi Yamashita Go version in AlphaZero 2017 finished the training in 34 hours according to Table S3. And it looks like AlphaZero Symmetries in AlphaZero 2018 finished the training in the same time according to Figure S1. So I think that the authors had adopted AlphaZero Symmetries in 2017 paper by mistake and retried the experiment again in 2018 paper. In order to compensate symmetries with real self-plays, they generated 8 times more games and reduced positions per game to 1/8. It is just my guess^^ ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AlphaZero paper difference between 2017 and 2018
Hi, Number of learned positions from a game record pos steps minibatch games AlphaGoZero 293 ( 700,000 * 2048) / 4,900,000 3 days AlphaGoZero 219 (3,100,000 * 2048) / 29,000,000256 x 40 block, 40 days AlphaZero 2017 137 ( 700,000 * 4096) / 21,000,000 AlphaZero 201820 ( 700,000 * 4096) / 140,000,000 ELF 2019 154 (1,500,000 * 2048) / 20,000,000 AlphaZero(Chess) 65 ( 700,000 * 4096) / 44,000,000 AlphaZero(Shogi) 119 ( 700,000 * 4096) / 24,000,000 All Network is 256 x 20 blocks, except AlphaGoZero 40 days. Average of game moves are Go220 Chess 80 Shogi 120 So I had thought learning all positions(from a game) once is nice. But AlphaZero2018 uses only 20 positions from a game. By the way, I did not received any mails since Ingo's mail(Mar 1 2019). Erik reported in Feb 17 2019, It looks like gmail is broken again for this list. I never got Remi's Remi also reported in Mar 24 2019. (I found this from archives.) I have just found out that the list is not sending emails to my free.fr Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AlphaZero paper difference between 2017 and 2018
Hi, I found AlphaZero paper Table S3 is different 2017 and 2018. 2017 2018 Mini-batches 700k 700k Training Time 34h 13d Training Games 21 million 140 million Thinking Time 800 sims, 200ms 800 sims, 200ms Training Time is 34h ->13d 9.2 times Training Games is 21 million -> 140 million 6.6 times Chess and Shogi is same. And Figure 1 is also a bit different in Shogi and Go. Chess looks same. Why these numbers are so different? Is it typo? AlphaZero(2017/12/05) Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm https://arxiv.org/abs/1712.01815 AlphaZero(2018/12/07) A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play https://deepmind.com/documents/260/alphazero_preprint.pdf Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] A new ELF OpenGo bot and analysis of historical Go games
Hi Ingo, * How strong is the new ELF bot in comparison with Leela-Zero? from CGOS BayesElo, new ELF(ELFv2) is about +100 stronger than Leela-Zero. Rating Network size(Resnet layers x filters) LZ_05db_ELFv2_p800 355420x256 LZ_d13c_ELFv1_p800 353020x224 second release ELF LZ_62b541_ELF_p800 350020x224 first release ELF LZ_204_05d1_p400343340x256 latest(2019-02-13) LZ p800 means 800 playout/move. p400 means 800 playout/move. Leela Zero's playout is half. Because its net size is double. http://www.yss-aya.com/cgos/19x19/bayes.html Thanks, Hiroshi Yamashita On 2019/02/17 1:29, "Ingo Althöfer" wrote: Hi Remi, thanks you for the link. A few questions (to all who know something): * How strong is the new ELF bot in comparison with Leela-Zero? * How were komi values taken into account when analysing old go games with help of ELF? * How often does ELF propose moves played by AlphaGo (for instance in the games with Fan Hui, Lee Sedol, and in the sixty games from December 2017)? * Does ELF understand that the strength of AlphaGo increased from October 2015 to May 2017? Cheers, Ingo. ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] AI Ryusei 2018 result
Hi, Our program Natsukaze also used Leela Zero recent 70 selfplay games to train DNN. Ladder escape moves(4% of total games) are removed, and chasing not ladder(0.3%) also removed. But its DNN policy was weak, around CGOS 2100. Maybe it is because current LZ selfplay use t=1 not first 30 moves but all moves. I did not know this. I think this makes selfplay weaker +1000 Elo. Switch to t=1 for all self-play moves, i.e., randomcnt=999 https://github.com/gcp/leela-zero-server/pull/81 Thanks, Hiroshi Yamashita On 2018/12/19 2:01, Gian-Carlo Pascutto wrote: On 17/12/18 01:53, Hiroshi Yamashita wrote: Hi, AI Ryusei 2018 was held on 15,16th December in Nihon-kiin, Japan. 14 programs played preliminary swiss 7 round, and top 6 programs played round-robin final. Then, Golaxy won. Result https://www.igoshogi.net/ai_ryusei/01/en/result.html It appears the 2nd place finisher after Golaxy was a hybrid of Rn and Leela Zero, using rollouts to compensate for Leela's network being trained with the "wrong" komi for this competition: https://github.com/zakki/Ray/issues/171#issuecomment-447637052 https://img.igoshogi.net/ai_ryusei/01/data/11.pdf ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Computers wins against pros in 9x9
Dear Ingo, Yesterday, I also could not see archives. But it is ok now. http://computer-go.org/pipermail/computer-go/ Another archives was available yesterday. https://www.mail-archive.com/computer-go@computer-go.org/maillist.html Thanks, Hiroshi Yamashita On 2018/12/18 20:19, Xavier Combelle wrote: computer-go archive and the server is up for me Le 17/12/2018 à 20:07, "Ingo Althöfer" a écrit : Dear Hiroshi, thanks for the information. It seems that the archive of computer-go is down, or is it even the whole server? Cheers, Ingo. ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Computers wins against pros in 9x9
Hi, Two Japanese Pros played 10 games against two computers in 9x9. And Pros lost by 2 wins, 6 losses, 2 draws. Two Pros are Hirofumi Ohashi 6dan and Nobuaki Anzai 7dan. Two computers are Natsukaze and Kishin they are winner and runners-up in CGF open 2018. Natsukaze result from human side game 1 Black:Anzai win (on time) game 2 Black:Ohashi loss W+2 game 3 White:Anzai loss B+2 game 4 White:Ohashi win W+R game 5 Black:Anzai loss W+4 2 wins 3 losses Kishin game 1 Black:Anzai draw game 2 Black:Ohashi loss W+R game 3 White:Anzai draw game 4 White:Ohashi loss B+2 game 5 White:Ohashi loss B+2 3 losses 2 draws SGF http://www.yss-aya.com/20180728sgf.zip Komi 7.0, Chinese rule Time is human : 5 minutes + add 5 sec/move. (fishcer rule) programs: 450 seconds This event was broadcasted by Igo-Shogi Channel last month. Validation! AI 9x9 http://www.igoshogi.net/igo/special/ai9roban.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AI Ryusei 2018 result
Hi, AI Ryusei 2018 was held on 15,16th December in Nihon-kiin, Japan. 14 programs played preliminary swiss 7 round, and top 6 programs played round-robin final. Then, Golaxy won. Result https://www.igoshogi.net/ai_ryusei/01/en/result.html BSK got creative award because its DNN was learned from author's own 5000 games. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] GoGui 1.5.0
Hi Remi, Thank you for gogui update! after 10 years? Gomoku and renju support sounds good. Thanks, Hiroshi Yamashita On 2018/11/17 5:57, Rémi Coulom wrote: Hi, In case anybody is interested, we have released a new version of GoGui: https://github.com/Remi-Coulom/gogui/releases The main purpose of this version is to add support of other games via an extension of the GTP protocol: https://www.kayufu.com/gogui/rules.html It also has minor improvements that may be useful for Go programmers: - high-resolution icons for high-dpi screens - wait a little for program output before displaying a popup dialog It also incorporates improvements by lemonsqueeze. In particular - handicap support in gogui-twogtp - correct scoring of handicap games The release page has a Windows installer. For other platforms, you will have to compile from source. The root of the repository has an "ubuntu_setup.sh" that should compile everything provided you have a jdk and ant installed. The contents of that file should give you indications of what to do on other platforms: https://github.com/Remi-Coulom/gogui/blob/master/ubuntu_setup.sh I use this version for my gomoku, renju, and Othello programs. We decided to distribute it, as it might be useful to others. Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] GokuiNano is very strong on CGOS 9x9
Hi, GokuiNano is very strong 9x9 program on CGOS. http://www.yss-aya.com/cgos/9x9/cross/GokuiNano.html Maybe this is the strongest 9x9 program as far as I know. It plays Tengen(5,5) well. http://www.yss-aya.com/cgos/viewer.cgi?9x9/SGF/2018/09/01/581667.sgf I have thought Tengen was not good move anymore, but it may be wrong. 9x9 is still diffcult game. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGF Open 2018 was held
Hi, CGF Open was held July 21st and 22nd on The University of Electro-Communications in Chofu, Japan. There were 10 program in 9x9, and 10 programs in 19x19. 9x9 was round-robin in preliminary league, and final is top 6 programs round-robin. 19x19 was 5 round swiss. 9x9 1st Natsukaze 10-0 2nd Kishin 8-2 3rd Rn 6-4 19x19 1st Natsukaze 5-0 2nd Rn 4-1 3rd Ray3-2 Natsukaze is a successor of Aya by Tomoyuki Kaneko and me. 9x9 1st and 2nd program played some games with Japanese professional Hirofumi Ohashi 6d and Nobuaki Anzai 7d. This will be broadcasted by Igo-Shogi Channel around November. I will post the result after broadcast. CGF Open 2018 (all are in Japanese) http://hp.vector.co.jp/authors/VA012620/cgf2018/index.html Rule http://hp.vector.co.jp/authors/VA012620/cgf2018/cgf2018.html Participants http://hp.vector.co.jp/authors/VA012620/cgf2018/list2018.html result http://hp.vector.co.jp/authors/VA012620/cgf2018/result2018.html Game records SGF(9x9) http://hp.vector.co.jp/authors/VA012620/cgf2018/cgf2018_9x9.zip Game records SGF(19x19) http://hp.vector.co.jp/authors/VA012620/cgf2018/cgf2018_19x19.zip Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGOS 9x9 draw rate
Hi, I have checked CGOS 9x9 black winrate and draw rate. CGOS 9x9 black winrate http://www.yss-aya.com/cgos2018_03_07_black_winrate.png CGOS 9x9 draw rate http://www.yss-aya.com/cgos2018_03_07_draw_rate.png Following table means 1800 range(both player are (1800 <= rate < 1900)) games are 397, and black winrate is 0.460. I did not plot 1900 - 2100 because few games. It seems black winrate tends to down for strong players. Draw rate clearly tends to up. Maybe over 4300, all games become draw? CGOS from 2018/03/01 to 2018/07/31, total 130922 games rate games B wins Draw W wins Black winrate(draw is 0.5) Draw rate 1800: 397 =( 153/ 59/ 185) 0.4600.149 1900: 3 =(2/1/0) 0.8330.333 2000: 1 =(0/0/1) 0.0000.000 2100: 6 =(5/0/1) 0.8330.000 2200:81 =( 34/3/ 44) 0.4380.037 2300: 1259 =( 587/ 77/ 595) 0.4970.061 2400: 4923 =( 2346/ 409/ 2168) 0.5180.083 2500: 945 =( 391/ 128/ 426) 0.4810.135 2600: 107 =( 43/ 24/ 40) 0.5140.224 2700: 1490 =( 518/ 253/ 719) 0.4330.170 2800: 3354 =( 970/ 1135/ 1249) 0.4580.338 2900: 1375 =( 416/ 390/ 569) 0.4440.284 3000: 181 =( 60/ 31/ 90) 0.4170.171 3100:41 =( 11/ 16/ 14) 0.4630.390 Following are Black first move winrate. It seems Black tengen (5,5) is not best move anymore for players >= 2800. Maybe one reason is Leela Zero 9x9(LZ20_128_65_p3000) only plays (4,5) and never play (5,5). Black first move winrate both player < 2800 both player >= 2800 games black winrate games black winrate (5,5) 41929(59.2) 0.554 5072(36.7) 0.398 (4,4) 11349(16.0) 0.548 2510(18.2) 0.388 (4,5) 7676(10.8) 0.538 5407(39.1) 0.519 (3,4) 2777( 3.9) 0.239604( 4.4) 0.424 (3,3) 1402( 2.0) 0.212 87( 0.6) 0.282 (3,5) 1353( 1.9) 0.180111( 0.8) 0.500 -- total 70769 13815 http://www.yss-aya.com/cgos/9x9/standings.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Tencent World AI Weiqi Competition will be held in June 23 and 24
Hi, After swiss 7R, result are follows. Top 8 programs will play round-robin on FoxGo next month. Fin Lee ELF Gol AQ Oct Nor BAD Dol Ray Aya bye Wins Sol Rank 1 FineArt 1 1 1 1 1 1 1 7-0 27 1 2 LeelaZero 01 1 1 1 1 1 6-1 29 2 3 ELF00 1 0 1 1 1 4-3 30 3 4 Golaxy 00 0 1 1 1 1 4-3 29 4 5 AQ 00 1 0 1 1 1 4-3 26 5 6 Octopus00 0 1 1 1 1 4-3 24 6 7 Northern 00 0 1 1 0 1 3-4 25 7 8 BADUKi 0 0 0 0 1 1 1 3-4 22 8 9 DolBaram0 0 0 0 1 1 1 3-4 18 9 10 Raynz 00 0 0 0 1 1 2-5 24 10 11 Aya 0 0 1 0 0 0 1 2-5 19 11 Note: I'm not sure Chinense Program English name. Final result (AQ author, Yu Yamaguchi's twitter) https://pbs.twimg.com/media/Dgclm35VAAAgsEZ.jpg:large Game records can be seen on FoxGo. (;GM[1]SZ[19] PB[FineArt] PW[LeelaZero] DT[2018-06-23] RE[B+R]KM[7.5]RU[Chinese]TM[40minutes + 10x30sec] ;B[dp];W[dd];B[pq];W[pd];B[po];W[pn];B[qn];W[qm];B[pm];W[on] ;B[om];W[qo];B[rn];W[nn];B[ro];W[nm];B[ql];W[cn];B[cl];W[cf] ;B[co];W[dn];B[fp];W[hq];B[cg];W[dg];B[dh];W[ch];B[bg];W[bf] ;B[eg];W[df];B[ci];W[bh];B[bi];W[ag];B[em];W[en];B[hp];W[dl] ;B[iq];W[ck];B[dk];W[bl];B[el];W[cm];B[dj];W[eh];B[fi];W[gn] ;B[gq];W[hl];B[ei];W[di];B[hc];W[hi];B[dh];W[id];B[fh];W[ic] ;B[ec];W[dc];B[fe];W[gg];B[ef];W[qf];B[pb];W[nb];B[qc];W[qd] ;B[oc];W[pc];B[ob];W[qb];B[md];W[lc];B[ld];W[kc];B[od];W[io] ;B[ko];W[km];B[mh];W[kg];B[of];W[qh];B[lk];W[fk];B[pg];W[ri] ;B[lm];W[ln];B[kn];W[jm];B[ll];W[ol];B[pl];W[nk];B[jk];W[ij] ;B[mn];W[pk];B[qk];W[pj];B[gl];W[gk];B[ek];W[nc];B[li];W[ho] ;B[mp];W[fc];B[he];W[mg];B[ng];W[mf];B[hk];W[ik];B[hm];W[il] ;B[gm];W[jn];B[ie];W[hd];B[jh];W[ne];B[jo];W[in];B[gp];W[fo] ;B[nd];W[je];B[jf];W[ke];B[qj];W[pi];B[oe];W[np];B[nq];W[ip] ;B[rc];W[rb];B[re];W[rd];B[qe];W[rf];B[pe];W[sc];B[jq]) Thanks, Hiroshi Yamashita On 2018/06/22 5:32, "Ingo Althöfer" wrote: Dear Hiroshi, thank you for spreading the news. Please, keep us informed about the results on this weekend. (As sign of thankyou I will press my thumbs for the Japanese soccer team in the Wordchampionships.) Cheers, Ingo. Gesendet: Donnerstag, 21. Juni 2018 um 20:33 Uhr Von: "Hiroshi Yamashita" An: computer-go@computer-go.org Betreff: [Computer-go] Tencent World AI Weiqi Competition will be held in June 23 and 24 Hi, Tencent World AI Weiqi Competition will be held in June 23 and 24. 11 programs will play. 4 from China. Fine Art, and so on. 3 from Japan. AQ, Raynz, Aya. 2 from Korea. DolBaram, Baduki. 1 from Belgium. Leela Zero 1 from America. ELF OpenGo Top 8 programs will play round-robin on FoxGo next month. Top 4 programs will play final late July. Tencent World AI Weiqi Competition (in Chinese) http://sports.sina.com.cn/go/2018-06-21/doc-ihefphqk9706598.shtml 2018 Tencent World AI Weiqi Competition (in Chinese) https://txwq.qq.com/act/game/index.html Rule http://weiqi.qq.com/special/109.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Tencent World AI Weiqi Competition will be held in June 23 and 24
Hi, Tencent World AI Weiqi Competition will be held in June 23 and 24. 11 programs will play. 4 from China. Fine Art, and so on. 3 from Japan. AQ, Raynz, Aya. 2 from Korea. DolBaram, Baduki. 1 from Belgium. Leela Zero 1 from America. ELF OpenGo Top 8 programs will play round-robin on FoxGo next month. Top 4 programs will play final late July. Tencent World AI Weiqi Competition (in Chinese) http://sports.sina.com.cn/go/2018-06-21/doc-ihefphqk9706598.shtml 2018 Tencent World AI Weiqi Competition (in Chinese) https://txwq.qq.com/act/game/index.html Rule http://weiqi.qq.com/special/109.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGOS Real-time game viewer
Hi, CGOS 19x19 Real-time game viewer is available. https://deepleela.com/cgos Thank you for author of DeepLeela site. DeepLeela logins as a Viewer, and dispatches to thier clients. There were Viewing Client on Linux and Windows, but no on the browser. Source code is also available. https://github.com/deepleela Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Using 9x9 policy on 13x13 and 19x19
I found a bug... In 9_i50_F64L11_p.prototxt, "policy head", name: "conv_1x1_2_policy" num_output: 2 should be name: "conv_1x1_1_policy" num_output: 1 Hiroshi Yamashita On 2018/03/02 4:21, Hiroshi Yamashita wrote: Hi Imran, I have only changed input shape. On Caffe, from 9x9 input_dim: 1 input_dim: 50 input_dim: 9 input_dim: 9 to 19x19 input_dim: 1 input_dim: 50 input_dim: 19 input_dim: 19 This is available on fully convolutional. http://computer-go.org/pipermail/computer-go/2015-December/008324.html Caffe's prototxt is http://www.yss-aya.com/20180302using9x9_19x19.tar.gz Thanks, Hiroshi Yamashita On 2018/03/02 0:35, Imran Hendley wrote: Hi Hiroshi, Are you using zero-padding to allow input shapes to match for all board sizes? Thanks, Imran ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Using 9x9 policy on 13x13 and 19x19
Hi Imran, I have only changed input shape. On Caffe, from 9x9 input_dim: 1 input_dim: 50 input_dim: 9 input_dim: 9 to 19x19 input_dim: 1 input_dim: 50 input_dim: 19 input_dim: 19 This is available on fully convolutional. http://computer-go.org/pipermail/computer-go/2015-December/008324.html Caffe's prototxt is http://www.yss-aya.com/20180302using9x9_19x19.tar.gz Thanks, Hiroshi Yamashita On 2018/03/02 0:35, Imran Hendley wrote: Hi Hiroshi, Are you using zero-padding to allow input shapes to match for all board sizes? Thanks, Imran ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Crazy Stone is back
Hi Remi, Wow, "Weights" is your engine. So my guess was right :-) In 9x9 CGOS, did you train in 9x9, or just use 19x19 network? Weights_33_400 is stronger than Weights_40_400. Maybe it is because Weights_33_400 use CrazyStone's playout, and Weights_40_400 does not use? Thanks, Hiroshi Yamashita On 2018/02/28 15:13, Rémi Coulom wrote: Hi, I have just connected the newest version of Crazy Stone to CGOS. It is based on the AlphaZero approach. The "Weights" engine were in fact previous experimental versions. CrazyStone-18.03 is using time control and pondering instead of a fixed number of evaluations per move. So it should be much stronger than Weights_31_3200. Does anybody know who cronus is? It is _extremely_ strong. Its rating is low because it has had only weaker opponents, but it is undefeated so far, except for one loss on time, and some losses against other versions of itself. It has just won two games in a row against Crazy Stone. I hope the other strong engines will reconnect, too. Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Crazy Stone is back
No Zen on CGOS is pity. To LZ-0xx-p1600-t1-r1 author, I think LZ-073-p1600-t1-r1 has BayesElo already. From LeelaZero page, 73 2018-02-05 23:0654bfb7b8 LZ-54bfb7-t1-p1600, BayesElo is 2903. Recalculating CGOS rating is not essential. And too many same kind bots running makes many selfplay matching. Its rating is more unreliable. Could you stop them, and run up to two or three bots which has no BayesElo? Thanks, Hiroshi Yamashita On 2018/02/28 17:12, Hideki Kato wrote: Welcome back Remi! On the 19x19 cgos, recently many LeelaZeros are running. This flood is making CGOS less useful and so I'll reconnect Zen after the flooding ends. Sorry for inconvinience. Hideki ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] 9x9 is last frontier?
Hi, Top 19x19 program reaches 4200 BayesElo on CGOS. But 3100 in 9x9. Maybe it is because people don't have much interest in 9x9. But it seems value network does not work well in 9x9. Weights_33_400 is maybe made by selfplay network. But it is 2946 in 9x9. Weights_31_3200 is 4069 in 19x19 though. In year 2012, Zen played 6 games against 3 Japanese Pros, and lost by 0-6. And it seems Zen's 9x9 strength does not change big even now. http://computer-go.org/pipermail/computer-go/2012-November/005556.html I feel there is still enough chance that human can beat best program in 9x9. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Using 9x9 policy on 13x13 and 19x19
Hi, Using 19x19 policy on 9x9 and 13x13 is effective. But opposite is? I made 9x9 policy from Aya's 10k playout/move selfplay. Using 9x9 policy on 13x13 and 19x19 19x19 DCNNAyaF128from9x91799 13x13 DCNNAyaF128from9x91900 9x9 DCNN_AyaF128a558x12290 Using 19x19 policy on 9x9 and 13x13 19x19 DCNN_AyaF128a523x12345 13x13 DCNNAya795F128a5232354 9x9 DCNN_AyaF128a523x12179 19x19 policy is similar strength on 13x13 and 166 Elo weaker on 9x9. 9x9 policy is 390 Elo weaker on 13x13, and 491 Elo weaker on 19x19. It seems smaller board is more useless than bigger board... Note: All programs select maximum policy without search. All programs use opening book. 19x19 policy is Filter128, Layer 12, without Batch Normalization. 9x9 policy is Filter128, Layer 11, without Batch Normalization. 19x19 policy is made from pro 78000 games, GoGoD. 9x9 policy is made from 10k/move. It is CGOS 2892(Aya797c_p1v1_10k). Ratings are BayesElo. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CLeelaz9Test on CGOS 9x9
Hi, It seems CLeelaz9Test rating has changed from 1700 to 2500. http://www.yss-aya.com/cgos/9x9/cross/CLeelaz9Test.html Could CLeelaz9Test author change its name when program get stronger? like CLeelaz9_1 or CLeelaz9_2. CGOS users may want to know correct rating quickly. Same program name but different strength is not good for CGOS rating system. Especially, BayesElo assumes all program strength are fixed. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Aya's selfplay games for 9x9 and 13x13
Hi, I have uploaded Aya's selfplay games for 9x9 and 13x13. 9x9 2,850,000 games 1. select random U from 1 to 104. 2. played up to U by Policy network probability. 3. U+1 move is played completely random. ";W[gg]C[RANDOM]" in sgf means this. 4. played with 1 playouts/move to the last. 5. Komi 7.0, CGOS rate is Aya794a523_ro_10k, 2529(BayesElo). 9x9_10k_r104_144x20k https://drive.google.com/drive/folders/1wVm-MOYDYXiQRMYWTqBM-R40jTXK1F-9?usp=sharing 13x13 8,640,000 games 1. played up to 16 moves by Policy network probability. 2. played with 2000 playouts/move to the last. 5. Komi 7.5, CGOS rate is Aya795b_523_ro_2k, 2603(BayesElo). 13x13_2k_r16_75_432x20k https://drive.google.com/drive/folders/1abpjezLGiifzTYR-AUxcIT0PjvCUv5_W?usp=sharing 19x19 games has moved to Google drive. https://drive.google.com/drive/folders/0B-qjJNUdVi15amRWSVRaZjRwNWc?usp=sharing Aya's selfplay games for training value network http://www.yss-aya.com/ayaself/ayaself.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] guess AlphaGo Zero's CGOS rating
Hi, I guessed AlphaGo Zero's CGOS rating. I picked up ratings from "Mastering the Game of Go without Human Knowledge" Figure 3 and Figure 6. GoRatings Raw network3030 (40 blocks) (Policy) AlphaGo Zero 5200 (40 blocks) 1 machine, 4 TPU AlphaGo Master 4800 1 machine, 4 TPU, Rollout AlphaGo Zero 4400 (20 blocks) 1 machine, 4 TPU AlphaGo Lee3700 many machines, 48 TPU 2016/03 AlphaGo Fan3150 many machines, 176 GPU 2015/10 CrazyStone 1900 Pachi 1300 GNU Go 400 I think over AlphaGo Lee, ratings calculated from selfplay. So I estimate rating progress is half. For example, Zero(5200) is +1500 stronger than Lee(3700). But I estimate Zero is +750 stronger ,Zero(4450). BayesElo CGOSGoRatings AlphaGo Zero 5422? 4450? (40 blocks) 1 machine, 4 TPU AlphaGo Master 5231? 4250? 1 machine, 4 TPU, Rollout AlphaGo Zero 5022? 4050? (20 blocks) 1 machine, 4 TPU AlphaGo Lee4672? 3700 many machines, 48 TPU 2016/03 Park Junghwan 4592? 3620 strongest human Ke Jie 4590? 3618 Yuta Iyama 4546? 3574 Lee Sedol 4514? 3542 Zen-15.7-4c1g 42693297? CGOS strongest Cho Chikun 4187? 3215 AlphaGo Fan4122? 3150 many machines, 176 GPU 2015/10 tianrang-galaxy40623090? Raw network4002? 3030 (40 blocks) (Policy) Fan Hui3978? 3006 AQ-2.1.1-4t1g 39442972? AlphaGo(Fan) 3862? 2890 1 machine(48CPU, 8GPU) 2015/10 Rn.4.16-4c 35252553? Aya792p2v2cn50_12t 33132341? KGS 7d (Blitz) AlphaGo RL 2729? 1757? Reinforcement learning from AlphaGo SL Aya790e_510_ro_1k 25801608? KGS 3d AlphaGo SL 2489? 1517 (Policy) Filter 192 pachi11_12t_100k 24291457? pachi, 100k playout/move Gnugo-3.7.10-a11800 828? KGS 5k Note 1. CGOS <-> GoRatings is based on AlphaGo RL beats Pachi 100k with 85%(+300). Note 2. CGOS = GoRatings + 972 Note 3. "?" is guess. Note 4. AlphaGo Lee(3700) may be calculated from 4-1 result against Lee Sedol(3542). Note 5. Human ratings are from https://www.goratings.org/en/ Note 6. CGOS rating is BayesElo. http://www.yss-aya.com/cgos/19x19/bayes.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AQ is updated
Hi, Latest AQ is updated. Author Yu Yamaguchi says this is AI Ryusei-sen version in 2017 December. CGOS is 3950(BayesElo). https://github.com/ymgaq/AQ/releases https://twitter.com/ymg_aq/status/955424462693834752 The latest version (v2.1.1) of Go program AQ is released! Update neural network files (vl.pb/sl.pb). Fix bugs of Nakade and Seki. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Japanese Shogi scene and AlphaZero
Hi Ingo, In Shogi, AlphaZero is +400 Elo stronger than elmo. But current strongest Shogi program is +184 stronger than elmo. AlphaZero is still strongest thought. Shogi open source program rating http://www.uuunuuun.com/english One slight things, "elmo" author Takizawa likes to be called "elmo", not "Elmo". Because Elmo is famous Sesame Street character. Like CGOS, many Shogi programs play on net. http://wdoor.c.u-tokyo.ac.jp/shogi/logs/LATEST/players-floodgate.html http://wdoor.c.u-tokyo.ac.jp/shogi/floodgate.html (Japanese) I also want to see AlphaZero Shogi game records. Thanks, Hiroshi Yamashita - Original Message - From: ""Ingo Althöfer"" <3-hirn-ver...@gmx.de> To: <computer-go@computer-go.org> Sent: Saturday, December 16, 2017 4:29 PM Subject: [Computer-go] Japanese Shogi scene and AlphaZero Hello in the round, I asked a Japanese friend how the DeepMind report (on Chess and Shogi) was recepted in the Shogi scene. Here are the central parts of his answer (given on Wednesday): Today, I just found a news on Alpha Zero in a Japanese common newspaper, Asahi Shinbun, reporting the superiority to previous AI playing program, probably through Google. ... In the paper, “Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorith,” by DeepMind, it seems there is no example of Shogi play. Japan Shogi League seems not yet to respond to recent Alpha Zero news. For the Shogi scene it would definitely be interesting to get Shogi games of AlphaZero for replay. Ingo. ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] kgs-chat command not working
Hi, I'm trying to implement the kgs-chat command but I'm getting a crash in I think kgs-chat command does not work since kgsGtp 3.5.20. Not supporting kgs-chat command may be safe. [Computer-go] kgsGTP client http://computer-go.org/pipermail/computer-go/2015-September/007914.html [Computer-go] kgsGtp 3.5.20 has a problem? http://computer-go.org/pipermail/computer-go/2015-October/007975.html Thanks, Hiroshi Yamashita - Original Message - From: "Andy" <andy.olsen...@gmail.com> To: "computer-go" <computer-go@computer-go.org> Sent: Monday, December 11, 2017 7:19 AM Subject: [Computer-go] kgs-chat command not working I'm trying to implement the kgs-chat command but I'm getting a crash in kgsGtp-3.5.22. Everything works fine and I can chat with the bot. But when a match starts that's when it crashes. Here is my gtp log. Any ideas? Or can someone paste me an example log that works or point me to example code? ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] AI ryusei 2017 first day result
Hi Remi, FineArt lost against Maru by Japanese rule. FineArt did not pass after Maru's pass. Tianrang also lost one game for this. Thanks, Hiroshi Yamashita - Original Message - From: "Rémi Coulom" <remi.cou...@free.fr> To: <computer-go@computer-go.org> Sent: Sunday, December 10, 2017 4:35 AM Subject: Re: [Computer-go] AI ryusei 2017 first day result Thanks Hiroshi. Did anything special happen in the game between Maru and FineArt? I wish you good games for the second day. Rémi ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AI ryusei 2017 first day result
Hi, 18 programs played swiss 7R, 1. DeepZenGo 6-1 2. FineArt6-1 3. DolBaram 5-2 4. Tianrang 5-2 5. Aya5-2 6. AQ 5-2 7. Maru 4-3 8. Abacus 4-3 9. Deep_ark 4-3 10. SR Go 3-4 Category B 11. Raynz 3-4 12. nlp3-4 13. GNU Go 3-4 Guest 14. Kugutsu2-5 15. Katsunari 2-5 16. Kifuwarabe 2-5 Category B 17. KinoaIgo 1-6 18. MayouiGo 0-7 15 programs will play tommorow tornament. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm
Hi, DeepMind makes strongest Chess and Shogi programs with AlphaGo Zero method. Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm https://arxiv.org/pdf/1712.01815.pdf AlphaZero(Chess) outperformed Stockfish after 4 hours, AlphaZero(Shogi) outperformed elmo after 2 hours. Search is MCTS. AlphaZero(Chess) searches 80,000 positions/sec. Stockfishsearches 70,000,000 positions/sec. AlphaZero(Shogi) searches 40,000 positions/sec. elmo searches 35,000,000 positions/sec. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AI RYUSEI 2017, new Computer Go tournament
Hi, AI RYUSEI 2017, new Computer Go tournament will be held in December 9 and 10 in Akihabara, Tokyo, Japan. This is a successor of UEC Cup. First day is preliminary league, swiss 7R. Second day is tournament for top 16 programs. AI RYUSEI 2017 http://www.igoshogi.net/ai_ryusei/01/en/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] November KGS bot tournament
Hi Nick, this will be the last of the series of KGS bot tournaments. Thank you for holding KGS tournament since 2005. On CGOS, there are always some new comers. I hope they also enter KGS bot tournament. Thanks, Hiroshi Yamashita - Original Message - From: "Nick Wedd" <mapr...@gmail.com> To: <computer-go@computer-go.org> Sent: Thursday, October 26, 2017 4:43 PM Subject: [Computer-go] November KGS bot tournament The November KGS bot tournament will be on Sunday, November 5th, starting at 16:00 UTC and ending by 22:00 UTC. It will use 19x19 boards, with time limits of 14 minutes each and very fast Canadian overtime, and komi of 7½. It will be a Swiss tournament. See http://www.gokgs.com/tournInfo.jsp?id=112 <http://www.gokgs.com/tournInfo.jsp?id=1116>7 Please register by emailing me at mapr...@gmail.com, with the words "KGS Tournament Registration" in the email title. With the falling interest in these events since the advent of AlphaGo, it is likely that this will be the last of the series of KGS bot tournaments. Nick -- Nick Wedd mapr...@gmail.com ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] AlphaGo Zero
I have two questions. 2017 Jan, Master , defeat 60 pros in a row. 2017 May, Master?, defeat Ke Jie 3-0. Master is Zero method with rollout. Zero is Zero method without rollout. Did AlphaGo that played with Ke Jie use rollout? Is Zero with rollout stronger than Zero without rollout? Thanks, Hiroshi Yamashita - Original Message - From: <cazen...@ai.univ-paris8.fr> To: <computer-go@computer-go.org> Sent: Thursday, October 19, 2017 2:50 AM Subject: [Computer-go] AlphaGo Zero https://deepmind.com/blog/ http://www.nature.com/nature/index.html Impressive! ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AQ is open source
Hi, AQ's author Yu Yamaguchi published his strong Go program AQ. https://github.com/ymgaq/AQ AQ went to 4th in 2017 UEC cup, and best 8 in 1st World AI Go Open. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGOS server update
Hi,
Recently CGOS 19x19 stopped often.
I checked log, and found it stopped
in "proc infoMsg {msg}" function, when it sends message to all.
-
error writing "sock18": connection reset by peer
while executing
"puts $soc "info $msg""
(procedure "infoMsg" line 8)
-
So I modified from
puts $soc "info $msg"
to
catch {puts $soc "info $msg"}
and restarted.
For viewer code, "catch" has been used, like
catch {puts $v "info $msg"}
I hope this change makes server stable.
Current server code is
http://www.yss-aya.com/20170914cgos_yama.zip
cgos.tcl... server
webuild.tcl ... making html
viewer.cgi ... sgf viewer
My change from original code.
-
webuild.tcl
recent 300 games
viewer.cgi with wgo.
cgos.tcl
forbidden "1024" login name (numeric only account)
handling draw
catch {puts $soc "info $msg"} in proc infoMsg {msg}.
sometimes CGOS server -socket function stopped here.
-
Thanks,
Hiroshi Yamashita
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGI won against FineArt in preliminary league.
(unofficial) Final result can be seen https://twitter.com/UgenNihonkiin/status/898109924604919808 Thanks, Hiroshi Yamashita - Original Message - From: ""Ingo Althöfer"" <3-hirn-ver...@gmx.de> To: <computer-go@computer-go.org> Sent: Thursday, August 17, 2017 7:08 PM Subject: Re: [Computer-go] CGI won against FineArt in preliminary league. Dear Hiroshi, thank you for the results and the sgf data. Congratulations to Zen team, including Hideki! Cheers, Ingo. ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGI won against FineArt in preliminary league.
Zen won. Final game and semi final "Zen vs FineArt" sgf are, (;GM[1]SZ[19] PB[CGI] PW[DeepZenGo] DT[2017-08-17]RE[W+R]KM[7.5]RU[Chinese] ;B[pq];W[pd];B[cp];W[dd];B[qf];W[nc];B[md];W[mc];B[ql];W[ep] ;B[gq];W[cq];B[jp];W[cn];B[bq];W[dq];B[do];W[bp];B[bo];W[co] ;B[ap];W[dp];B[cm];W[bn];B[bm];W[dn];B[dm];W[an];B[cf];W[fd] ;B[fm];W[oo];B[np];W[no];B[mp];W[ol];B[pj];W[qm];B[rm];W[rn] ;B[qn];W[pm];B[ro];W[rl];B[sn];W[qk];B[qj];W[rj];B[ri];W[nj] ;B[sj];W[rk];B[pk];W[pl];B[nh];W[mo];B[pn];W[on];B[nm];W[nl] ;B[mm];W[po];B[qp];W[oi];B[qh];W[lp];B[om];W[oq];B[km];W[ll] ;B[lm];W[pr];B[qr];W[kq];B[lj];W[mi];B[or];W[qo];B[rn];W[rp] ;B[op];W[ko];B[sl];W[pp];B[qq];W[lk];B[jk];W[in];B[kj];W[lh] ;B[jo];W[jn];B[kn];W[hp];B[kp];W[lo];B[jq];W[nq];B[mq];W[jr] ;B[lq];W[kr];B[nn];W[lr];B[mr];W[ms];B[nr];W[iq];B[mn];W[pi] ;B[qi];W[jl];B[il];W[ik];B[ij];W[hk];B[kl];W[hl];B[fo];W[en] ;B[fn];W[em];B[el];W[fl];B[eo];W[bp];B[hr];W[ir];B[cp];W[io] ;B[ln];W[bp];B[ek];W[cp];B[gl];W[hj];B[fi];W[dg];B[ii];W[hi] ;B[di];W[fp];B[ee];W[de];B[nf];W[ih];B[df];W[ef];B[eg];W[ff] ;B[fg];W[lf];B[kg];W[lg];B[lc];W[ld];B[nd];W[ne];B[gf];W[fe] ;B[le];W[kd];B[od];W[oc];B[ke];W[me];B[oe];W[mf];B[je];W[kc] ;B[of];W[jg];B[lb];W[kb];B[hg];W[if];B[ge];W[ie];B[gd];W[hc] ;B[cc];W[bd];B[bf];W[gc];B[cd];W[dc];B[cb];W[rd];B[qc];W[qd] ;B[db];W[eb];B[pc];W[pb];B[qb];W[ob];B[rb];W[rc];B[sb];W[ra] ;B[re];W[qa];B[jm];W[am];B[bk];W[ak];B[aj];W[bl];B[cl];W[al] ;B[cj];W[dk];B[dj];W[oj];B[ip];W[gp];B[sk];W[ql];B[qe];W[sc] ;B[hq];W[fq];B[fb];W[ea];B[pe];W[sa];B[hd];W[id];B[fc];W[ed] ;B[hm];W[hh];B[ho];W[hn];B[ic];W[ga];B[mj];W[ni];B[jd];W[ib] ;B[ce];W[ec];B[hb];W[gb];B[jb];W[jc];B[fr];W[hs];B[ka];W[ja] ;B[kh];W[kf];B[jh];W[gh];B[gg];W[go];B[gr];W[er]) (;GM[1]SZ[19] PB[FineArt] PW[DeepZenGo] DT[2017-08-17]RE[W+R]KM[7.5]RU[Chinese] ;B[pd];W[dp];B[qp];W[dc];B[fq];W[cn];B[lq];W[pm];B[op];W[qj] ;B[qh];W[kp];B[de];W[kq];B[lp];W[lo];B[ko];W[mo];B[np];W[jo] ;B[kn];W[ip];B[dr];W[cq];B[ep];W[gp];B[do];W[eo];B[co];W[dn] ;B[bp];W[bo];B[cp];W[bq];B[dq];W[ap];B[dp];W[bm];B[hr];W[kl] ;B[ec];W[eb];B[ed];W[fc];B[fb];W[gb];B[db];W[fa];B[cc];W[qf] ;B[pf];W[pe];B[qe];W[oe];B[qd];W[pg];B[rf];W[qg];B[rg];W[ph] ;B[ge];W[od];B[hc];W[rh];B[cj];W[ob];B[pb];W[ej];B[ic];W[cb] ;B[da];W[cd];B[dd];W[bc];B[dc];W[ch];B[bg];W[bh];B[cg];W[dh] ;B[fo];W[en];B[jn];W[in];B[lm];W[ll];B[mm];W[nn];B[jm];W[ik] ;B[jk];W[jl];B[im];W[hn];B[il];W[hk];B[gm];W[fn];B[ml];W[po] ;B[pp];W[lj];B[mj];W[mi];B[mk];W[li];B[pk];W[qk];B[on];W[om] ;B[no];W[mn];B[jj];W[lk];B[gk];W[gj];B[ql];W[ok];B[fk];W[fj] ;B[oj];W[pj];B[ni];W[oi];B[nj];W[nm];B[nh];W[mg];B[pl];W[rl] ;B[of];W[og];B[mh];W[lh];B[lg];W[ng];B[ji];W[jh];B[hj];W[hi] ;B[ij];W[ih];B[hl];W[fl];B[ek];W[dk];B[el];W[fm];B[dj];W[cl] ;B[gn];W[go];B[fp];W[iq];B[bk];W[fh];B[gq];W[hq];B[dl];W[ck] ;B[bl];W[cm];B[ho];W[hp];B[kr];W[jq];B[bj];W[an];B[gi];W[gh] ;B[ii];W[hh];B[ei];W[fi];B[al];W[am];B[ar];W[aq];B[br];W[qm] ;B[oc];W[nc];B[pc];W[aj];B[nd];W[bi];B[ne];W[lf];B[kg];W[kh] ;B[kf];W[ke];B[mf];W[jf];B[le]) Hiroshi Yamashita - Original Message - From: ""Ingo Althöfer"" <3-hirn-ver...@gmx.de> To: <computer-go@computer-go.org> Sent: Thursday, August 17, 2017 5:02 PM Subject: Re: [Computer-go] CGI won against FineArt in preliminary league. Dear Hiroshi, thank you for our information. Can you please keep us informed about the results from the final tournament? Thank you in advance, Ingo. Gesendet: Donnerstag, 17. August 2017 um 02:59 Uhr Von: "Hiroshi Yamashita" <y...@bd.mbn.or.jp> An: computer-go@computer-go.org Betreff: [Computer-go] CGI won against FineArt in preliminary league. Hi, CGI won against FineArt and Zen in preliminary league. https://twitter.com/UgenNihonkiin/status/897771608743399424 Top 8 programs will play today's final tournamet. World AI Go Open https://www.reddit.com/r/baduk/comments/6ge1ul/the_1st_world_ai_go_open/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] CGI won against FineArt in preliminary league.
Hi Ingo, Zen beated FineArt in semi final. Final is CGI vs Zen. Live is available in FoxGo. Thanks, Hiroshi Yamashita - Original Message - From: ""Ingo Althöfer"" <3-hirn-ver...@gmx.de> To: <computer-go@computer-go.org> Sent: Thursday, August 17, 2017 5:02 PM Subject: Re: [Computer-go] CGI won against FineArt in preliminary league. Dear Hiroshi, thank you for our information. Can you please keep us informed about the results from the final tournament? Thank you in advance, Ingo. Gesendet: Donnerstag, 17. August 2017 um 02:59 Uhr Von: "Hiroshi Yamashita" <y...@bd.mbn.or.jp> An: computer-go@computer-go.org Betreff: [Computer-go] CGI won against FineArt in preliminary league. Hi, CGI won against FineArt and Zen in preliminary league. https://twitter.com/UgenNihonkiin/status/897771608743399424 Top 8 programs will play today's final tournamet. World AI Go Open https://www.reddit.com/r/baduk/comments/6ge1ul/the_1st_world_ai_go_open/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] CGI won against FineArt in preliminary league.
Hi, CGI won against FineArt and Zen in preliminary league. https://twitter.com/UgenNihonkiin/status/897771608743399424 Top 8 programs will play today's final tournamet. World AI Go Open https://www.reddit.com/r/baduk/comments/6ge1ul/the_1st_world_ai_go_open/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] successor of UEC Cup will be held in December 9, 2017
Web page says, - The successor tournament of the UEC Cup will be held on the following schedule. Date: 9th to 10th December 2017 The organizer of this new event is "IGO & SHOGI CHANNEL INC.". The regulation and rule of the competition will be planned to inherit past UEC-cup. Details such as the venue will be announced in July. (June 06, 2017) http://www.computer-go.jp/uec/public_html/eng/index.shtml - Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Multi-Labelled Value Networks for Computer Go
Hi, CGI team published interesting paper. Their value network is +130 Elo stronger than simple value network. Multi-Labelled Value Networks for Computer Go https://arxiv.org/abs/1705.10701 By the way, I had not received any mails since May 23. And today I got three Kato's mail that posted May 24. From archives, Jan van der Steen also reported same trouble. http://computer-go.org/pipermail/computer-go/2017-May/010125.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AlphaGo won second game against Ke Jie, 2-0
(;GM[1]SZ[19] PB[AlphaGo] PW[Ke Jie] DT[2017-05-25] RE[B+R] KM[7.5]TM[180]RU[Chinese]PC[Wuzhen, China]GN[game2] ;B[qp];W[pd];B[cq];W[cd];B[ec];W[oq];B[pn];W[df];B[nc];W[qf] ;B[pc];W[qc];B[qb];W[oc];B[pb];W[od];B[ob];W[rc];B[nd];W[mb] ;B[lc];W[lb];B[qd];W[rd];B[jc];W[mc];B[pe];W[oe];B[ld];W[kp] ;B[iq];W[fp];B[dn];W[io];B[ch];W[cl];B[eh];W[cg];B[bg];W[bf] ;B[fn];W[em];B[en];W[ek];B[kq];W[lq];B[lp];W[jq];B[kr];W[jp] ;B[jr];W[mq];B[ip];W[ho];B[gp];W[dq];B[go];W[cp];B[lo];W[kn] ;B[ln];W[km];B[dp];W[hp];B[hq];W[gq];B[gr];W[fq];B[hn];W[jo] ;B[fr];W[do];B[co];W[ep];B[bp];W[md];B[ne];W[of];B[lm];W[le] ;B[kl];W[jl];B[gl];W[kk];B[ll];W[jk];B[mj];W[gj];B[pq];W[pr] ;B[qr];W[po];B[qo];W[on];B[op];W[np];B[pm];W[om];B[no];W[bo] ;B[dp];W[oo];B[pp];W[cp];B[cn];W[nl];B[lk];W[mo];B[pl];W[nj] ;B[ni];W[ok];B[nf];W[mi];B[li];W[mh];B[og];W[pf];B[ki];W[ij] ;B[lg];W[nh];B[oi];W[oh];B[pj];W[ph];B[pk];W[bq];B[dp];W[jg] ;B[kg];W[cp];B[jn];W[in];B[dp];W[fo];B[cp];W[gn];B[ke];W[me] ;B[jf];W[jb];B[ic];W[ib];B[er];W[hc];B[cc];W[bc];B[dd];W[cb] ;B[ce];W[dc];B[cf];W[dg];B[be]) Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] AlphaGo watch parties planned across U.S.
NicoNico movie shows Zen's evaluation. (In Japanese) http://live2.nicovideo.jp/watch/lv298444014 Zen thinks White(AlphaGo) 's winrate is 62% at 86th moves. Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] AlphaGo watch parties planned across U.S.
I'll sometimes post Aya's evaluation. https://twitter.com/nhk_igo_15bai Thanks, Hiroshi Yamashita - Original Message - From: "Horace Ho" <hora...@gmail.com> To: "computer-go" <computer-go@computer-go.org> Sent: Tuesday, May 23, 2017 10:47 AM Subject: Re: [Computer-go] AlphaGo watch parties planned across U.S. https://www.youtube.com/watch?v=Z-HL5nppBnM On Tue, May 23, 2017 at 8:27 AM, Hideki Kato <hideki_ka...@ybb.ne.jp> wrote: Live will start at 11:00 AM JST (GMT+0900). https://events.google.com/alphago2017/ Hideki ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] mini-max with Policy and Value network
Hi, HiraBot author reported mini-max search with Policy and Value network. It does not use monte-carlo. Only top 8 moves from Policy is searched in root node. In other depth, top 4 moves is searched. Game result against Policy network best move (without search) Win Loss winrate MaxDepth=1, (558-442) 0.558 +40 Elo MaxDepth=2, (351-150) 0.701 +148 Elo MaxDepth=3, (406-116) 0.778 +218 Elo MaxDepth=4, (670- 78) 0.896 +374 Elo MaxDepth=5, (490- 57) 0.896 +374 Elo MaxDepth=6, (520- 20) 0.963 +556 Elo Search is simple alpha-beta. There is a modification Policy network high probability moves tend to be selected. MaxDepth=6 takes one second/move on i7-4790k + GTX1060. His nega-max code http://kiyoshifk.dip.jp/kiyoshifk/apk/negamax.zip CGOS result, MaxDepth=6 http://www.yss-aya.com/cgos/19x19/cross/minimax-depth6.html His Policy network(without search) is maybe http://www.yss-aya.com/cgos/19x19/cross/DCNN-No336-tygem.html His Policy and Value network(MCTS) is maybe http://www.yss-aya.com/cgos/19x19/cross/Hiratuka10_38B100.html Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Ke Jie vs. AlphaGo match
DeepMind says we can see streaming live. AlphaGo at The Future of Go Summit, 23-27 May 2017 https://deepmind.com/research/alphago/alphago-china/ DeepMind @DeepMindAI 2h https://twitter.com/DeepMindAI/status/865518993695625217 Excited to see what we’ll discover at the Future of Go Summit next week. Stay informed on our new webpage #AlphaGo17 https://goo.gl/TzUvE6 Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Ke Jie vs. AlphaGo match
Is it japanese tome zone? I think it is Japanese time. So UTC is Ke Jie vs. AlphaGo (3 hours + 1 minute x5) Game1May 23 02:30-09:30 Game2May 25 02:30-09:30 Game3May 27 02:30-09:30 Pair Go May 26 00:30-03:30 Team Go May 26 04:30-10:30 Panda net site says game record live is available. Viewer soft is free, but I'm not sure whether guest account is ok. Ke Jie vs. AlphaGo Moves http://ugi.pandanet.co.jp/?key=alphago-2017-1 Moves with comment by Rin Kono 9 pro http://ugi.pandanet.co.jp/?key=alphago-2017-1c Pair Go Moves http://ugi.pandanet.co.jp/?key=alphago-2017-4 Moves with comment by Daisuke Murakawa 8 pro http://ugi.pandanet.co.jp/?key=alphago-2017-4c Team Go Moves http://ugi.pandanet.co.jp/?key=alphago-2017-5 Moves with comment by Daisuke Murakawa 8 pro http://ugi.pandanet.co.jp/?key=alphago-2017-5c Viewer for Windows http://www.pandanet.co.jp/setup/ Viewer for iPhone https://itunes.apple.com/jp/app/pandanet-go/id406456426 Viewer for Android https://play.google.com/store/apps/details?id=be.gentgo.tetsuki=search_result#?t=W251bGwsMSwxLDEsImJlLmdlbnRnby50ZXRzdWtpIl0. Thanks, Hiroshi Yamashita - Original Message - From: "Xavier Combelle" <xavier.combe...@gmail.com> To: <computer-go@computer-go.org> Sent: Friday, May 19, 2017 3:52 PM Subject: Re: [Computer-go] Ke Jie vs. AlphaGo match Is it japanese tome zone? ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Ke Jie vs. AlphaGo match
Hi, It will be played in a week. But there are few information about this. Is there YouTube live available? I found a schedule in Panda-net site. Ke Jie vs. AlphaGo (3 hours + 1 minute x5) Game1May 23 11:30-18:30 Game2May 25 11:30-18:30 Game3May 27 11:30-18:30 Pair Go May 26 09:30-12:30 Team Go May 26 13:30-19:30 Panda net (in Japanese) http://www.pandanet.co.jp/event/fogs/ Exploring the mysteries of Go with AlphaGo and China's top players https://deepmind.com/blog/exploring-mysteries-alphago/ Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] May KGS bot tournament
Hi Ingo, Maybe we can guess from CGOS BayesElo. http://www.yss-aya.com/cgos/19x19/bayes.html Zen-14.6-1c1g4169 Aya792p2v2cn50_12t 3472 AyaMC is similar to Aya792p2v2cn50_12t, around 3500. Zen19X uses 6 threads with one GPU. So maybe around 4169 + 250 = 4300. Zen19X is +800 Elo stronger, and its winrate is about 99%. Thanks, Hiroshi Yamashita - Original Message - From: ""Ingo Althöfer"" <3-hirn-ver...@gmx.de> To: <computer-go@computer-go.org> Sent: Tuesday, May 09, 2017 7:07 PM Subject: Re: [Computer-go] May KGS bot tournament Hi Hideki, thanks for the explannation. So, Aya was already too far ahead. Zen had two more games against Aya in the rounds 5-12 - and won both of them. What do you think about the playing strengths of Zen and Aya in comparison? Cheers, Ingo. ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Welcome to Download the "computer go dataset" (160K9D vs 9D SGFs)
Hi Yu Yuan, Thanks! I wrote a converter. http://www.yss-aya.com/20170418tygem_convert.tar.gz I could not convert some of Tygem 2011 and 2014, because of id mismatch. Thanks, Hiroshi Yamashita - Original Message - From: "Yu Yuan" <yen...@gmail.com> To: <computer-go@computer-go.org> Sent: Tuesday, April 18, 2017 12:19 AM Subject: [Computer-go] Welcome to Download the "computer go dataset" (160K9D vs 9D SGFs) 2015.11.02 - 2016.12.31 TYGEM "9D vs 9D" dataset (1,516,031 games). 2003.09.25 - 2011.12.28 TOM "9D vs 9D" dataset (50,956 games). https://github.com/yenw/computer-go-dataset <https://github.com/yenw> ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Google TPU performance
Hi, Table 1 "CNN0" is maybe AlphaGo? It is Conv Layers 16, Weights 8M. So maybe 256 Filters? Paper says TPU is 15X - 30X faster than GPU(K80) and CPU. On my CNN, TITAN X Pascal is 4.3X faster than K80. So maybe TPU is 3X - 4X faster than TITAN X Pascal. Thanks, Hiroshi Yamashita - Original Message - From: "Peter Kollarik" <peter.kolla...@gmail.com> To: "computer-go" <computer-go@computer-go.org> Sent: Thursday, April 06, 2017 7:32 AM Subject: [Computer-go] Google TPU performance i just found this: https://drive.google.com/file/d/0Bx4hafXDDq2EMzRNcy1vSUxtcEk/view Peter ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Aya's selfplay games for training value network
Hi, I have uploaded Aya's selfplay games. There are about 1,440,000 games. Selfplay program strength is about KGS 4d and 5d. Aya's selfplay games for training value network http://www.yss-aya.com/ayaself/ayaself.html Using these games, KGS 4d over, tygem 9d, and GoGoD, I made slightly (+70 Elo) better value network than using KGS 4d over, tygem 9d, and GoGoD. White tends to win in KGS komi 0.5 games. So I deleted 41% white win games in komi 0.5 game. Value network is 64 Filters, 14 Layers with batch normalization. Caffe's prototxt is http://www.yss-aya.com/ayaself/20170405aya_value.tar.gz A game has 4 parts. 1. First move is selected from opening book. And rorated 8 symmetry randomly. 2. 2nd - 16th moves are selected randomly from Policy network probability. 3. 2000 playouts/move self play. And it ends when a player resigns. 4. resume with 300 playouts/move selfplay from resign position. And remove all dead stones. This 300 playouts/move selfplay is about KGS 3d. Only root node is created by Policy network. Komi is 7.5, Chinese rule. Note: If Aya plays (1,1) or (1,19) or (19,1) or (19,19) or (9,1)(=J19) within 16 moves, it is bug. It should be deleted. If original game result "GN[B+R," and game result "RE[B+3.5]" is different (like RE[W+5.5]), it should be deleted. 300 playout/move makes mistakes. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Zen won against Ryo Ichiriki in Densei-sen
FineArt won against Ryo Ichiriki. (;GM[1]SZ[19] PB[FineArt] PW[Ryo Ichiriki] DT[2017-03-26] RE[B+R] KM[6.5]TM[60]RU[Japanese] PC[Bellesalle Iidabashi First, Tokyo] GN[Densei-sen 2017 2nd game] ;B[pd];W[dp];B[dc];W[qp];B[cn];W[fq];B[ch];W[jc];B[nc];W[gc] ;B[fd];W[eb];B[gd];W[ec];B[ed];W[hc];B[cc];W[oq];B[en];W[bp] ;B[gp];W[fp];B[gn];W[di];B[dh];W[ei];B[bj];W[hi];B[qo];W[qm] ;B[rp];W[qq];B[oo];W[mp];B[qj];W[nn];B[on];W[qk];B[pj];W[pk] ;B[nm];W[oj];B[ph];W[mn];B[ol];W[ok];B[mm];W[ln];B[rq];W[qr] ;B[rn];W[lk];B[lm];W[km];B[np];W[nq];B[kl];W[ll];B[kn];W[jm] ;B[lp];W[mo];B[mq];W[ko];B[jn];W[lq];B[mr];W[kq];B[jo];W[kp] ;B[kk];W[im];B[lj];W[mk];B[hl];W[pm];B[om];W[hm];B[gm];W[ik] ;B[mj];W[nj];B[ki];W[jk];B[nh];W[oh];B[og];W[mi];B[kh];W[rj] ;B[ri];W[mg];B[mh];W[qh];B[qi];W[rh];B[rk];W[li];B[kj];W[lh] ;B[ni];W[oi];B[lg];W[si];B[sj];W[sh];B[hn];W[rm];B[gj];W[hk] ;B[gk];W[ih];B[gi];W[if];B[ie];W[jf];B[jd];W[lf];B[kc];W[ic] ;B[hg];W[ig];B[hf];W[kg];B[ng];W[kd];B[ke];W[ld];B[je];W[le] ;B[lc];W[kf];B[he];W[nd];B[od];W[hj];B[hh];W[ii];B[jb];W[ib] ;B[kb];W[db];B[cb];W[da];B[ne];W[po];B[pn];W[qn];B[pp];W[pg] ;B[pf];W[rr];B[qg];W[pi];B[co];W[cp];B[gq]) Thanks, Hiroshi Yamashita - Original Message - From: "Hiroshi Yamashita" <y...@bd.mbn.or.jp> To: <computer-go@computer-go.org> Sent: Sunday, March 26, 2017 12:55 PM Subject: [Computer-go] Zen won against Ryo Ichiriki in Densei-sen Hi, Zen won against Ryo Ichiriki in Densei-sen. Second game Ryo Ichiriki vs Fine Art will be played in a hour. ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Zen won against Ryo Ichiriki in Densei-sen
Hi, Zen won against Ryo Ichiriki in Densei-sen. Second game Ryo Ichiriki vs Fine Art will be played in a hour. (;GM[1]SZ[19] PB[Ryo Ichiriki] PW[DeepZenGo] RE[W+R] DT[2017-03-26]KM[6.5]RU[Japanese] PC[Bellesalle Iidabashi First, Tokyo] GN[Densei-sen 2017 1st game] ;B[pd];W[pp];B[cd];W[dp];B[ic];W[qf];B[nd];W[qc];B[qd];W[rd] ;B[cn];W[fq];B[dj];W[rb];B[ed];W[cc];B[dc];W[ce];B[bc];W[dh] ;B[fi];W[fg];B[hh];W[bd];B[cb];W[cj];B[ck];W[ci];B[bk];W[dk] ;B[ej];W[dl];B[ep];W[cm];B[do];W[bm];B[eq];W[fp];B[fo];W[er] ;B[dq];W[hp];B[go];W[jq];B[qn];W[pn];B[po];W[qo];B[oo];W[qm] ;B[qp];W[rn];B[ro];W[qn];B[pq];W[rp];B[op];W[rq];B[lq];W[dr] ;B[cr];W[en];B[eo];W[br];B[cq];W[jd];B[jc];W[ld];B[pf];W[pg] ;B[og];W[of];B[nf];W[pe];B[oe];W[oh];B[pf];W[qe];B[ke];W[kd] ;B[he];W[if];B[kg];W[id];B[hd];W[hc];B[gc];W[hb];B[gb];W[kc] ;B[ib];W[jb];B[ha];W[nb];B[jr];W[ir];B[kr];W[in];B[ho];W[ip] ;B[hr];W[gr];B[is];W[hq];B[iq];W[io];B[gp];W[ir];B[ob];W[oc] ;B[iq];W[gq];B[nc];W[mb];B[pc];W[pb];B[od];W[oa];B[ph];W[of] ;B[qb];W[qa];B[pf];W[qg];B[oj];W[ng];B[of];W[nh];B[pi];W[mi] ;B[mg];W[nk];B[nj];W[mj];B[mk];W[lk];B[ml];W[kj];B[nl];W[jh] ;B[jg];W[ig];B[ih];W[kh];B[lh];W[ij];B[me];W[jf];B[kf];W[mh] ;B[le];W[li];B[lg];W[fk];B[gk];W[gl];B[hk];W[il];B[fl];W[gm] ;B[gg];W[bo]) Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] UEC cup 2nd day
FineArt won. (;GM[1]SZ[19] PB[zen] PW[fineart] DT[2017-03-19]RE[W+R]KM[6.5]TM[30]RU[Japanese]PC[UEC, Tokyo] ;B[qd];W[dc];B[pq];W[dp];B[oc];W[po];B[qo];W[qn];B[qp];W[pm] ;B[pj];W[oq];B[pp];W[op];B[oo];W[pn];B[no];W[or];B[pr];W[lq] ;B[lo];W[rn];B[kq];W[kr];B[mr];W[mq];B[lr];W[kp];B[jq];W[lp] ;B[jr];W[jp];B[hq];W[hp];B[gp];W[ho];B[fq];W[ml];B[fn];W[jl] ;B[cn];W[dn];B[dm];W[co];B[bn];W[bo];B[en];W[do];B[ce];W[ed] ;B[mk];W[ll];B[ol];W[om];B[nl];W[nm];B[ic];W[ph];B[og];W[pg] ;B[oh];W[pe];B[ne];W[qe];B[rd];W[qj];B[qk];W[qi];B[pi];W[rk] ;B[gc];W[df];B[cg];W[eh];B[di];W[cm];B[cl];W[bm];B[bl];W[an] ;B[jf];W[dl];B[dk];W[em];B[bd];W[bc];B[cc];W[cb];B[cd];W[fc] ;B[gb];W[gd];B[hd];W[ge];B[hf];W[rq];B[ps];W[pk];B[eb];W[db] ;B[bb];W[ba];B[ab];W[gh];B[fi];W[dg];B[fh];W[fg];B[dh];W[ig] ;B[if];W[gg];B[ii];W[gi];B[mi];W[ei];B[dj];W[ki];B[kj];W[ji] ;B[jj];W[ih];B[dr];W[kg];B[lf];W[nj];B[ni];W[re];B[fj];W[sd] ;B[rc];W[od];B[pd];W[oe];B[nd];W[nc];B[ob];W[rb];B[sc];W[nf] ;B[mf];W[qb];B[pc];W[nb];B[na];W[lc];B[of];W[lb];B[pf];W[rg] ;B[se];W[rf];B[ma];W[sb];B[oa];W[la];B[ld];W[kd];B[kc];W[jc] ;B[ke];W[jd];B[jb];W[kb];B[ja];W[id];B[hc];W[md];B[me];W[mb] ;B[mc];W[ib];B[ia];W[md];B[rh];W[pb];B[pa];W[qh];B[mc];W[qa] ;B[sa];W[md];B[qg];W[le];B[qf];W[ri];B[ld];W[cf];B[bf];W[le] ;B[sh];W[si];B[ld];W[he];B[ie];W[le]) - Original Message - From: "Hiroshi Yamashita" <y...@bd.mbn.or.jp> To: <computer-go@computer-go.org> Sent: Sunday, March 19, 2017 2:38 PM Subject: Re: [Computer-go] UEC cup 2nd day Final is Fine Art vs Zen ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] UEC cup 1st day result
Today, 30 participanst played, and 1st Fine Art 7-0 2nd Zen6-1 lost aaginst FineArt 3rd CrazyStone 6-1 4th Rayn 5th Aya5-2 6th AQ 7th CGI 8th Julie 9th DolBaram Top 16 program will play tommorow tournament. Remi's post is useful. https://twitter.com/Remi_Coulom/status/843115954473070593 There was a lot of media this year. Maybe because of "Fine Art" (very strong program made by Tencent) Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] UEC Cup broadcast info
Hi, Individual games can be watched by logging-in the NNGS server For holding the tournament safely, this is not available now. Sorry for inconvenience. Thanks, Hiroshi Yamashita - Original Message - From: "Hideki Kato" <hideki_ka...@ybb.ne.jp> To: <computer-go@computer-go.org> Sent: Friday, March 17, 2017 1:47 PM Subject: Re: [Computer-go] UEC Cup broadcast info Individual games can be watched by logging-in the NNGS server using some NNGS client program such as Jago. See official page http://www.computer-go.jp/uec/public_html/eng/network.shtml for the name of the server, which is global this year. NB: Please leave enough bandwidth for the entrants. Hideki ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] How is zen so strong on CGOS?
Hideki said Zen has used policy network since ver 10.5, and value network since ver 12.1. Maybe value network is very effective. Thanks, Hiroshi Yamashita - Original Message - From: "Detlef Schmicker" <d...@physik.de> To: <computer-go@computer-go.org> Sent: Wednesday, January 25, 2017 11:53 PM Subject: [Computer-go] How is zen so strong on CGOS? Hi, I'd like to start a discussion on what zen might do being so strong on CGOS with only one core and no graphic card :) The version actual playing (and therefore best comparable to the programs actual playing) is Rank Name Elo + − Games 12 Zen-13.5-1c0g 3639 48 48 569 16 Aya792p2v2cn50_12t 3558 46 46 476 24 Rn.3.3-4c 3459 34 34 918 26 No335-4.5-gpu-4c 3438 115 115 112 29 CGI1407_1_475_7c 3399 73 73 142 53 CrazyStone-0002 3239 114 114 141 54 NG-05 3235 67 67 228 No335 is probably Hirabot NG is oakfoam from crazystone I dont know, if this version uses gpu Rn is probably ray All but Zen use gpu and so have probably about 300 policy network calls per second. Zen does this on cpu and I got Hideki to let me know, that this can do only about 20 calls per second :) It might be that zen uses cnn only in the upper nodes and relies on the faster node generation zen used before cnn for deeper nodes? I can not believe, that the policy network is so much stronger than the ones of all the others?! Any ideas Detlef ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Golois5 is KGS 4d
Hi, Golois5 is KGS 4d. I think it is a first bot that gets 4d by using DCNN without search. Golois5 (4d) info says "I use a policy network trained on Gogod games." Golois4 (3d) info says "I use a policy network trained on KGS games played by 6 dan or more." If both network are same, GoGoD games is better than KGS games for DCNN? Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] it's alphago
Ray was Japanese student program that went on 7th, UEC cup 2016. Ray http://computer-go-ray.com/eng/index.html Thare is a stronger version of Ray, with policy net and value net. https://github.com/zakki/Ray/tree/nn CGOS BayesElo is 3463 (Rn.3.3-4c). http://www.yss-aya.com/cgos/19x19/bayes.html Hiroshi Yamashita - Original Message - From: "Andy" <andy.olsen...@gmail.com> To: "computer-go" <computer-go@computer-go.org> Sent: Friday, January 06, 2017 11:48 PM Subject: Re: [Computer-go] it's alphago What is Ray? Strongest open source bot? Anyone have a link to it? ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] it's alphago (How to get a strong value network)
Hi, If I understood correctly you would try to use a program using value net with (let's say 2000 playouts) in selfplay? Using only one result, or Yes. 2000 playouts/move MCTS with policy net and value net. doing some games per position? Or are you thinking of using only the win I thought one game per position, but some games per position looks nice option. doing some games per position? Or are you thinking of using only the win percentage such a program gives from his own mixing of SL network, In my experience, game result is better than win percentage. usage per game (at least if Rn3.3-4c is Ray on CGOS with 4 cores, NG04b Oh NG04b is oakfoam. AlphaGo RL is about 2800 (CGOS BayesElo). So around this rating program seems nice. And many computers don't have GPU. To calculate DCNN on CPU, maybe we can not use big network(filter 192), but smaller one(filter 64 or 32). Hiroshi Yamashita - Original Message - From: "Detlef Schmicker" <d...@physik.de> To: <computer-go@computer-go.org> Sent: Friday, January 06, 2017 7:20 PM Subject: Re: [Computer-go] it's alphago (How to get a strong value network) Hi, this sounds interesting! AlphaGo paper plays only with RL network, if I understood correctly. If we start this huge approach we should try to carefully discuss the way (and hopefully get some hints from people tried with much computational power :) If I understood correctly you would try to use a program using value net with (let's say 2000 playouts) in selfplay? Using only one result, or doing some games per position? Or are you thinking of using only the win percentage such a program gives from his own mixing of SL network, search and value net? By the way to make some promotion :) oakfoam is not far away from Ray for this kind of approach, where you will probably try to reduce cpu/gpu usage per game (at least if Rn3.3-4c is Ray on CGOS with 4 cores, NG04b is oakfoam on CGOS with 10k and saving GPU usage by using only 50% of GX970) Detlef Am 06.01.2017 um 10:39 schrieb Hiroshi Yamashita: If value net is the most important part for over pro level, the problem is making strong selfplay games. 1. make 30 million selfplay games. 2. make value net. 3. use this value net for selfplay program. 4. go to (1) I don't know when the progress will stop by this loop. But if once strong enough selfplay games are published, everyone can make pro level program. 30 million is big number. It needs many computers. Computer Go community may be able to share this work. I can offer Aya, it is not open-source though. Maybe Ray(strongest open source so far) is better choice. Thanks, Hiroshi Yamashita - Original Message - From: <fotl...@smart-games.com> To: <computer-go@computer-go.org> Sent: Friday, January 06, 2017 4:50 PM Subject: Re: [Computer-go] it's alphago Competitive with Alpha-go, one developer, not possible. I do think it is possible to make a pro level program with one person or a small team. Look at Deep Zen and Aya for example. I expect I’ll get there (pro level) with Many Faces as well. David ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Deep Zen vs Cho Chikun -- Round 3
Hi, Cho Chikun won game3 against DeepZenGo, and won the match. (;GM[1]SZ[19] PB[Cho Chikun] PW[DeepZenGo] DT[2016-11-23]RE[B+R]KM[6.5]TM[120]RU[Japanese] PC[Hotel New Otani]EV[2nd Igo DenOu-sen]GN[Game 3] ;B[pd];W[dp];B[pq];W[dd];B[fq];W[cn];B[pk];W[mp];B[jp];W[po] ;B[oo];W[on];B[op];W[pm];B[nn];W[nm];B[mm];W[nl];B[mo];W[ml] ;B[lm];W[qf];B[qh];W[of];B[oh];W[nd];B[rd];W[mg];B[nc];W[mc] ;B[oc];W[ni];B[md];W[ld];B[me];W[le];B[ne];W[oe];B[lc];W[mb] ;B[od];W[kc];B[nf];W[ng];B[og];W[mf];B[nd];W[dg];B[dc];W[cc] ;B[ec];W[cb];B[gc];W[hq];B[dr];W[jq];B[kq];W[kp];B[lp];W[ko] ;B[lq];W[ip];B[cq];W[eq];B[er];W[fp];B[ci];W[ch];B[di];W[fh] ;B[cl];W[qi];B[pi];W[qj];B[pj];W[rh];B[qk];W[rk];B[rg];W[qg] ;B[ph];W[ri];B[rf];W[ql];B[fe];W[fr];B[bo];W[co];B[bp];W[ic] ;B[ib];W[ej];B[bm];W[kl];B[km];W[ed];B[fd];W[fb];B[hc];W[id] ;B[eb];W[jb];B[eg];W[eh];B[df];W[cf];B[de];W[ce];B[ee];W[cd] ;B[fg];W[qp];B[qq];W[rq];B[gh];W[dh];B[rr];W[hh];B[gi];W[hg] ;B[gg];W[oa];B[bh];W[bg];B[ei];W[bi];B[fi];W[ah];B[jr];W[jo] ;B[ir];W[fm];B[qn];W[pn];B[rp];W[ro];B[sq];W[pb];B[rn];W[qc] ;B[qe];W[ij];B[ll];W[lk];B[sl];W[qo];B[so];W[rl];B[kk];W[lj] ;B[jl];W[kj];B[in];W[hn];B[nj];W[bn];B[an];W[db];B[ea];W[ln] ;B[mn];W[bk];B[mi];W[li];B[jj];W[ji];B[mk]) Japanese movie news Japanese Go AI lost against top pro by 1 win 2 losses http://www3.nhk.or.jp/news/html/20161123/k10010781571000.html https://www.youtube.com/watch?v=B0U6ZwyC1-0 Japanese newspapers http://www.yomiuri.co.jp/culture/20161123-OYT1T50036.html http://www.nikkei.com/article/DGXLASDG23H6A_T21C16A1CR8000/ http://www.nikkei.com/article/DGXLASFG23H0Y_T21C16A100/ http://mainichi.jp/graphs/20161123/hpj/00m/040/002000g/1 http://www.asahi.com/articles/ASJCR4GV5JCRUCVL004.html Regards, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Aya reaches pro level on GoQuest 9x9 and 13x13
Hi Detlef, You did not try reinforcement learning I think. Do you have any idea, why this would make the policy network 250ELO stronger, as mentioned in the alphago paper (80% winrate)? I have not tried reinforcement learning, but I guess if threre are two moves, SL probability are taking 5 stones(35%), good shape(37%). RL may change this taking 5 stones(80%), good shape(10%). For weaker player, taking 5 stones is maybe safe. Do you think playing strength would be better, if one only takes into account the moves of the winning player? I think learning only from winning player moves will get better result. Now I'm making 13x13 selfplay games like AlphaGo paper. 1. make a position by Policy(SL) probability from initial position. 2. play a move uniformly at random from available moves. 3. play left moves by Policy(RL) to the end. (2) means it plays very bad move usually. Maybe it is because making completely different position? I don't understand why this (2) is needed. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Aya reaches pro level on GoQuest 9x9 and 13x13
Hi, Did you not find a benefit from a larger value network? Too little data and too much overfitting? Or more benefit from more frequent evaluation? I did not find larger value network is better. But I think I need more taraining data and stronger selfplay. I did not find overfitting so far, and did not try more frequent evaluation. Policy + Value vs Policy, 1000 playouts/move, 1000 games. 9x9, komi 7.0 0.634 using game result. 0 or 1 I presume this is a winrate, but over what base? Policy network? Yes. Policy network(only root node) + value network vs Policy network(only root node). How do you handle handicap games? I see you excluded them from the KGS dataset. Can your value network deal with handicap? I excluded hadicap games. My value network can not handle hadicaps. It it only for komi 7.5. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] UEC Cup will be held in 18-19 March 2017
Hi, UEC Cup will be held in 18-19 March 2017. The UEC Cup : 18-19 March 2017 Registration deadline : 11 December 2016 DenSei-Sen: late March or April This is the last UEC Cup and DenSei-sen. The Tenth UEC Cup Computer Go http://www.computer-go.jp/uec/public_html/eng/index.shtml Regards, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Aya reaches pro level on GoQuest 9x9 and 13x13
lue net, CGOS 2900 https://github.com/zakki/Ray/tree/nn Dark Forest https://github.com/facebookresearch/darkforestGo DeltaGo, trace AlphaGo Policy SL. Accuracy 54%. DCNN is calculated on CPU. http://home.q00.itscom.net/otsuki/delta.html Regards, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Zen19K2 is strongest player on KGS
Hi, Zen19K2 is strongest player on KGS. http://www.gokgs.com/top100.jsp Oops, another player is top now. But anyway nearly top. Zen19K2 is maybe 10.3d from graph. http://www.gokgs.com/graphPage.jsp?user=Zen19K2 Zen19K2's information is - Computer program Zen running on KURISU server provided by DWANGO. CPU: Xeon E5-2623 v3 x2 GPU: GeForce GTX TITAN X x4 The number of handicap stones is limited to 3 or less. - Congratulations for graduation from KGS, Zen! I think Zen19K2 strength is similar to 2015/10 AlphaGo that beated Fan Hui 2p, 5-0. Thanks, Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] AlphaGo selfplay 3 games
Hi, DeepMind published AlphaGo's selfplay 3 games with comment. AlphaGo Games - English https://deepmind.com/research/alphago/alphago-games-english/ OHASHI Hirofumi 6d pro said "I understood AlphaGo is extremely strong even if 5sec/move." Thanks Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] cgos <--> kgs rating
Hi, I have some data. CGOS is 19x19. CGOSKGS Botname Time Aya788d_p1v0_6c12t 32356d AyaMC1min + 15sec x10 Aya786m_4c 29654d AyaMC4 1min + 15sec x10 AyaF128_507_ro_1k 27363d AyaBotD1 10min + 30sec x5 Aya786l_10k 26752d AyaMC2 10min + 30sec x5 Aya786m_5k 25642d AyaBot2 10min + 30sec x5 Aya786m_1k 21761k AyaBot01 10min + 30sec x5 Gnugo-3.7.10-a1 18005k AyaBot 13929k AyaBot 10min + 30sec x5 This post has also some info. guess AlphaGo's CGOS rating http://computer-go.org/pipermail/computer-go/2016-February/008638.html Thanks, Hiroshi Yamashita - Original Message - From: "Detlef Schmicker" <d...@physik.de> To: <computer-go@computer-go.org> Sent: Tuesday, June 28, 2016 3:05 AM Subject: [Computer-go] cgos <--> kgs rating Hay, I wonder if somebody has the same program (with the same settings) rated on cgos 19x19 and kgs? I am still fighting with resigning in the case of value-network and playouts disagree, so I can not run oakfoam on kgs, but would like to have a strength hint :) Thanks Detlef ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Densei-sen
Zen won against KOBAYASHI with 3 handicaps. (;GM[1]SZ[19] PB[Zen] PW[KOBAYASHI Koichi] DT[2016-03-23] RE[B+4.5]HA[3]KM[0.5]TM[1800]RU[Japanese] AB[pd][dp][pp] ;W[de];B[dc];W[cc];B[cb];W[cd];B[gc];W[cn];B[fq];W[cj];B[pj] ;W[nc];B[lc];W[qc];B[qd];W[pc];B[od];W[nd];B[oc];W[ob];B[ne] ;W[me];B[pb];W[mb];B[qb];W[rc];B[rb];W[oe];B[nf];W[rd];B[qe] ;W[re];B[qf];W[rf];B[qg];W[rg];B[qh];W[sb];B[md];W[lb];B[kc] ;W[kb];B[jc];W[nq];B[lq];W[mo];B[kp];W[pn];B[on];W[qq];B[pq] ;W[qp];B[po];W[qo];B[no];W[pm];B[om];W[pl];B[ol];W[jk];B[hk] ;W[jm];B[hm];W[ji];B[hi];W[jo];B[ln];W[jg];B[hg];W[mf];B[mg] ;W[ng];B[of];W[og];B[pf];W[mh];B[lg];W[lh];B[lf];W[ll];B[ch] ;W[eh];B[di];W[dk];B[ek];W[eb];B[dl];W[cl];B[dj];W[ck];B[ef] ;W[ee];B[ec];W[db];B[cf];W[df];B[dg];W[ff];B[eg];W[fd];B[fc] ;W[bb];B[bo];W[dm];B[el];W[bn];B[kn];W[jn];B[ip];W[hd];B[mj] ;W[mk];B[nk];W[lj];B[ml];W[lk];B[co];W[bh];B[bg];W[ci];B[dh] ;W[pr];B[or];W[qr];B[bi];W[bj];B[ah];W[do];B[eo];W[dn];B[cp] ;W[bl];B[nr];W[hc];B[jb];W[hb];B[qk];W[rl];B[ql];W[rm];B[qm] ;W[qn];B[mc];W[nb];B[ka];W[mm];B[gf];W[fe];B[fg];W[mn];B[mp] ;W[fo];B[en];W[hn];B[hp];W[fn];B[em];W[gm];B[nl];W[lm];B[nh] ;W[ni];B[oh];W[oi];B[pg];W[jp];B[jq];W[gl];B[nj];W[gk];B[gj] ;W[mi];B[kg];W[kh];B[jf];W[fp];B[ep];W[be];B[rk];W[ij];B[hj] ;W[gp];B[gq];W[rh];B[ia];W[ha];B[sl];W[rn];B[hl];W[fk];B[fj] ;W[jd];B[id];W[ie];B[ic];W[je];B[ho];W[im];B[la];W[ra];B[he] ;W[if];B[ma];W[qa];B[hf];W[gd];B[ig];W[kf];B[ke];W[kd];B[na] ;W[pa];B[le];W[jf];B[ih];W[jh];B[bf];W[oq];B[mr];W[np];B[mq] ;W[nn];B[op];W[ko];B[lo];W[km];B[lp];W[ri];B[io];W[in];B[fm] ;W[gn];B[ce];W[bd];B[il];W[jl];B[an];W[am];B[ao];W[aj];B[ai] ;W[os];B[ns];W[ps];B[pk];W[oj];B[ae];W[ad];B[af];W[pi];B[ok] ;W[iq];B[hq];W[rj];B[sj];W[sm];B[sk];W[si];B[ja];W[oa];B[ld]) Hiroshi Yamashita - Original Message - From: ""Ingo Althöfer"" <3-hirn-ver...@gmx.de> To: <computer-go@computer-go.org> Sent: Wednesday, March 23, 2016 5:10 PM Subject: Re: [Computer-go] Densei-sen Dear Hiroshi, many thanks for keeping us informed! Ingo. Gesendet: Mittwoch, 23. März 2016 um 08:06 Uhr Von: "Hiroshi Yamashita" <y...@bd.mbn.or.jp> An: computer-go@computer-go.org Betreff: Re: [Computer-go] Densei-sen SGF is here. (;GM[1]SZ[19] PB[darkforest] PW[KOBAYASHI Koichi] DT[2016-03-23] RE[W+R]HA[3]KM[0.5]TM[1800]RU[Japanese] AB[pd][dp][pp] ;W[de];B[dc];W[cc];B[ci];W[db];B[ec];W[cd];B[ic];W[cn];B[cl] ;W[fq];B[dn];W[cq];B[dq];W[cp];B[co];W[bo];B[do];W[bn];B[cm] ;W[dr];B[er];W[cr];B[eq];W[jp];B[lq];W[fo];B[fm];W[qf];B[nd] ;W[pj];B[ql];W[nq];B[lo];W[qq];B[pq];W[qp];B[po];W[qo];B[pn] ;W[pr];B[or];W[qr];B[jn];W[fr];B[fs];W[kq];B[lr];W[gs];B[es] ;W[hr];B[hn];W[hp];B[ph];W[qd];B[qc];W[qh];B[qi];W[pi];B[qj] ;W[pe];B[oh];W[pc];B[od];W[rc];B[qb];W[rb];B[pb];W[rh];B[mh] ;W[kj];B[kh];W[lm];B[kl];W[ll];B[kk];W[lk];B[jj];W[kn];B[ko] ;W[jm];B[jo];W[ni];B[nh];W[qk];B[pk];W[ji];B[ki];W[lj];B[ij] ;W[rk];B[nj];W[ok];B[pl];W[oj];B[nk];W[ol];B[nl];W[om];B[nm] ;W[on];B[nn];W[oo];B[no];W[op];B[oq];W[np];B[nr];W[mp];B[lp] ;W[eh];B[fe];W[dj];B[cj];W[hk];B[hj];W[fl];B[gm];W[gj];B[gi] ;W[fj];B[il];W[gh];B[hh];W[gg];B[fi];W[ei];B[dg];W[hg];B[ef] ;W[he];B[gc];W[dk];B[el];W[dl];B[dm];W[ek];B[gk];W[em];B[en] ;W[id];B[jc];W[gd];B[fd];W[ck];B[bk];W[hc];B[hb];W[gb];B[hd] ;W[bl];B[bm];W[hc];B[eb];W[fc];B[cg];W[bj];B[al];W[ee];B[ff] ;W[ib];B[jb];W[ha];B[fb];W[bf];B[if];W[ge];B[gc];W[jd];B[lc] ;W[bh];B[ch];W[bi];B[bg];W[ag];B[ah];W[ai];B[cb];W[bb];B[da] ;W[ba];B[ca];W[cf];B[fa];W[fn];B[el];W[fc];B[gl];W[ed];B[gc] ;W[ld];B[kc];W[fc];B[fk];W[eg];B[df];W[gc];B[ej];W[di];B[aj] ;W[ak];B[fh];W[fg];B[aj];W[mc];B[md];W[ak];B[dd];W[ce];B[aj] ;W[kd];B[ah];W[mb];B[lb];W[ka];B[la];W[ja];B[nc];W[ma];B[jr] ;W[ir];B[is];W[kr];B[gq];W[gp];B[ks];W[jq];B[ip];W[hs];B[qn] ;W[af]) Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
Re: [Computer-go] Densei-sen
SGF is here. (;GM[1]SZ[19] PB[darkforest] PW[KOBAYASHI Koichi] DT[2016-03-23] RE[W+R]HA[3]KM[0.5]TM[1800]RU[Japanese] AB[pd][dp][pp] ;W[de];B[dc];W[cc];B[ci];W[db];B[ec];W[cd];B[ic];W[cn];B[cl] ;W[fq];B[dn];W[cq];B[dq];W[cp];B[co];W[bo];B[do];W[bn];B[cm] ;W[dr];B[er];W[cr];B[eq];W[jp];B[lq];W[fo];B[fm];W[qf];B[nd] ;W[pj];B[ql];W[nq];B[lo];W[qq];B[pq];W[qp];B[po];W[qo];B[pn] ;W[pr];B[or];W[qr];B[jn];W[fr];B[fs];W[kq];B[lr];W[gs];B[es] ;W[hr];B[hn];W[hp];B[ph];W[qd];B[qc];W[qh];B[qi];W[pi];B[qj] ;W[pe];B[oh];W[pc];B[od];W[rc];B[qb];W[rb];B[pb];W[rh];B[mh] ;W[kj];B[kh];W[lm];B[kl];W[ll];B[kk];W[lk];B[jj];W[kn];B[ko] ;W[jm];B[jo];W[ni];B[nh];W[qk];B[pk];W[ji];B[ki];W[lj];B[ij] ;W[rk];B[nj];W[ok];B[pl];W[oj];B[nk];W[ol];B[nl];W[om];B[nm] ;W[on];B[nn];W[oo];B[no];W[op];B[oq];W[np];B[nr];W[mp];B[lp] ;W[eh];B[fe];W[dj];B[cj];W[hk];B[hj];W[fl];B[gm];W[gj];B[gi] ;W[fj];B[il];W[gh];B[hh];W[gg];B[fi];W[ei];B[dg];W[hg];B[ef] ;W[he];B[gc];W[dk];B[el];W[dl];B[dm];W[ek];B[gk];W[em];B[en] ;W[id];B[jc];W[gd];B[fd];W[ck];B[bk];W[hc];B[hb];W[gb];B[hd] ;W[bl];B[bm];W[hc];B[eb];W[fc];B[cg];W[bj];B[al];W[ee];B[ff] ;W[ib];B[jb];W[ha];B[fb];W[bf];B[if];W[ge];B[gc];W[jd];B[lc] ;W[bh];B[ch];W[bi];B[bg];W[ag];B[ah];W[ai];B[cb];W[bb];B[da] ;W[ba];B[ca];W[cf];B[fa];W[fn];B[el];W[fc];B[gl];W[ed];B[gc] ;W[ld];B[kc];W[fc];B[fk];W[eg];B[df];W[gc];B[ej];W[di];B[aj] ;W[ak];B[fh];W[fg];B[aj];W[mc];B[md];W[ak];B[dd];W[ce];B[aj] ;W[kd];B[ah];W[mb];B[lb];W[ka];B[la];W[ja];B[nc];W[ma];B[jr] ;W[ir];B[is];W[kr];B[gq];W[gp];B[ks];W[jq];B[ip];W[hs];B[qn] ;W[af]) Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
[Computer-go] Densei-sen
Hi, darkforest lost against Koichi Kobayashi with 3 handicaps. Next game, Zen vs Kobayashi will be played also with 3 handicaps. Hiroshi Yamashita ___ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go
