Jeudi 29 octobre 2009 4 29 /10 /Oct /2009 15:49
99a
The first ever win of a computer as black in 9x9 Go against a 9P.


Our Taiwanese partners for a ANR grant have organized public demonstration games between

MoGoTW (based on MoGo 4.86.Soissons + the "TW" modifications developped
jointly with our Taiwanese colleagues), with support by Grid5000
and
C.-H. Chou 9P, top pro player winner of the LG Cup 2007.


The algorithm is presented in another page http://teytaud.over-blog.com/article-35709049.html .
This was during a press conference at Taipei around a French-Taiwanese grant for joint research.

Comments by the pro:
He told the jounalists that he made a mistake on move 26 - an error hard to read, even for a professional Go player (see picture below).


99b.jpg
The suboptimal move made by the pro (move 24, on the left, center).


However, he was so surprised that MoGoTW found this mistake that he made during the game. That means the level of MoGoTW in 9x9 game has made an improvement than last time (08/2009).

In addition, he also told the journalists that MoGoTW made a good move at Move 29 (see picture below). He thought that the main reason that MoGoTW won the game was Move 29 and MoGoTW didn't make any mistake after Move 29.

99c.jpg
Very good move by mogoTW (move 29, on the upper right part).


The final position is indicated below; the big white group F6 is dead.



Authors of MoGo:
Contributors include all the mogo-people, plus Mei-Hui Wang, Chang-Shing Lee, Shi-Jim Yen, Hsien-Der Huang and Cheng-Wei Chou.

Details:
a) MoGoTW was running on 32 eight-cores in Taiwan.
b) There were two blitz games (15 minutes per side), won by the pro.
c) There was one non-blitz game (45 minutes per side). MoGo was unlucky
as it was black, but it nonetheless won the game. This game is enclosed.
All games can be found on KGS (account nutngo)



Remarks:


a) Fuego won as white against a 9P a few months ago. Therefore computers
have won both as white and black against top players :-) We now should
win on a complete game like 4 out of 7 games and the job would be
completly done for 9x9 Go :-)

b) MoGo already won a game as black, against Catalin Taranu, but I guess
the pro, at that time, had played an original opening somehow for fun
(I'm not sure of that, however).

c) My feeling is that blitz games are not favorable to computers... Statistics
are in accordance with this I guess. Humans are stronger for short time
settings.

d) The game was longer than most important games I've seen (59 moves).

All comments welcome.

Best regards
Olivier


============= the SGF file =========================================
(;GM[1]FF[4]CA[UTF-8]AP[CGoban:3]ST[2]
RU[Chinese]SZ[9]KM[7.50]TM[2700]
PW[nutngo]PB[MoGoTW9X9]DT[2009-10-26]PC[The KGS Go Server at http://www.gokgs.com/]C[MoGoTW9X9 [-\]: GTP Engine for MoGoTW9X9 (black): MoGo version MoGo version 4.86.Soissons: I'll resign if you win and playing until you pass if I win. If you passed and I did not, it often means that the status of some stones are not clear enough for me (I want to be sure we will agree on dead stones). Clarify them. Maybe I am losing, but have not enough time to be sure I should resign. In this case play until the end :'( (very rare cases)
]RE[B+Resign]
;B[ee]BL[2699.494]
;W[ge]WL[2652.472]
;B[gf]BL[2699.208]
;W[ce]WL[2597.429]
;B[cd]BL[2698.918]
;W[ff]WL[2574.488]
;B[fe]BL[2698.64]
;W[gg]WL[2573.551]
;B[hf]BL[2698.336]
;W[gd]WL[2572.792]
;B[fg]BL[2490.513]
;W[ef]WL[2569.373]
;B[de]BL[2283.362]
;W[fh]WL[2451.256]
;B[eg]BL[2092.766]
;W[df]WL[2436.298]
;B[gh]BL[1919.531]
;W[dg]WL[2431.654]
;B[hg]BL[1757.395]
;W[bd]WL[2340.808]
;B[cc]BL[1756.26]
;W[ec]WL[2283.551]
;B[be]BL[1755.969]
;W[cf]WL[2281.403]
;B[bf]BL[1755.694]
;W[eh]WL[2271.753]
;B[dd]BL[1610.253]
;W[bg]WL[2268.966]
;B[gc]BL[1476.576]
;W[fd]WL[2260.802]
;B[hc]BL[1354.659]
;W[fc]WL[2151.486]
;B[db]BL[1243.165]
;W[hd]WL[2141.275]
;B[gb]BL[1140.924]
;W[fa]WL[2119.7]
;B[id]BL[1047.266]
;W[ie]WL[2112.227]
;B[he]BL[961.171]
;W[ic]WL[2109.32]
;B[ib]BL[882.168]
;W[ha]WL[2104.112]
;B[id]BL[809.979]
;W[bc]WL[2100.375]
;B[bb]BL[743.745]
;W[ic]WL[2097.445]
;B[fb]BL[683.217]
;W[eb]WL[2094.669]
;B[id]BL[626.972]
;W[af]WL[2082.434]
;B[ga]BL[576.013]
;W[ab]WL[2068.078]
;B[ea]BL[529.084]
;W[ba]WL[2060.281]
;B[cb]BL[486.46]
;W[gg]WL[2057.484]
;B[hi]BL[447.483]
;W[fg]WL[2053.419]
;B[ih]BL[411.532])




Par teytaud - Publié dans : Recherche - Communauté : Science & Avenir
Ecrire un commentaire - Voir les 3 commentaires
Retour à l'accueil
Créer un blog gratuit sur over-blog.com - Contact - C.G.U. - Rémunération en droits d'auteur - Signaler un abus - Articles les plus commentés