Files
msproj/data/games/pro9x9/game011.sgf
Woody Folsom aca8320600 First working implementation of ANN which is trained using GameResults.
The PassFilter simply outputs BlackWins and WhiteWins (Range 0 - 1 but not presently clamped).
In principle, this type of feedforward ANN can be used to decide whether a PASS will result in blackwins or whitewins at any stage.
The goal is for the network to learn that passing while losing when valid moves exist is bad, but passing while winning is relatively harmless later in the game.
2012-11-17 18:40:31 -05:00

388 lines
1.3 KiB
Plaintext

(
;
GaMe[1]
VieW[]
SiZe[9]
KoMi[5.5]
EVent[TV game]
PlayerBlack[Aragaki Shumu (6 dan)]
PlayerWhite[Ishida Atsushi (5 dan)]
DaTe[1992]
PlaCe[Osaka]
REsult[B+4.5]
Comment[
Black: Aragaki Shumu (6 dan)
White: Ishida Atsushi (5 dan)
]
;
Black[ff]
;
White[fc]
;
Black[ce]
;
White[cf]
;
Black[de]
;
White[df]
;
Black[ef]
;
White[be]
;
Black[bd]
;
White[bf]
;
Black[dc]
;
White[bc]
;
Black[gd]
;
White[cc]
;
Black[db]
;
White[hd]
;
Black[he]
;
White[gc]
;
Black[fd]
;
White[hb]
;
Black[id]
;
White[ic]
;
Black[hc]
;
White[ed]
;
Black[ec]
;
White[hd]
;
Black[ie]
;
White[ge]
;
Black[fe]
;
White[gf]
;
Black[gg]
;
White[hf]
;
Black[hc]
;
White[cd]
;
Black[ib]
;
White[dd]
;
Black[fb]
;
White[eh]
;
Black[fh]
;
White[eg]
;
Black[fg]
;
White[cb]
;
Black[ei]
;
White[da]
;
Black[gb]
;
White[dh]
;
Black[di]
;
White[ci]
;
Black[fi]
;
White[ch]
;
Black[ea]
;
White[ee]
;
Black[ca]
;
White[ba]
;
Black[da]
C[Black wins by 4 1/2 point]
;
)