[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Bot Comparison after 60/500 games... GNU odd/even ply cube effects

Posted By: MaX
Date: Friday, 30 October 2009, at 4:24 p.m.

In Response To: Bot Comparison after 60/500 games... GNU odd/even ply cube effects (neilkaz)

MaX "the idea of supervised training where 0ply is trained to mimic 2ply (instead of 1ply) is a bad idea, at the origin of the odd-even crap."

Neil replies... is this how GNU was trained? I am unaware of how GNU was trained to be honest, but this odd/even ply stuff can be severe. This effect is clearly less in other bots.

I'm not 100% sure but I think that part of the supervised training was done to try to "force" 0ply to be as good as 2ply (i.e. rewarding the 0ply net when it made the same play as 2ply). I know very little in NN area, but this seems odd to me: I would have forced 0ply to mimic 1ply (first, at least).

MaX.

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.