[Computer-go] Congratulations to CrazyStone!
ds2 at physik.de
Fri Mar 8 09:42:59 PST 2013
Can you provide a link to your thesis, as the one I found is dead:)
Am Freitag, den 08.03.2013, 00:30 +0000 schrieb Aja Huang:
> Now it seems to me that this is related to the way playouts
> are done
> and it will be difficult to improve with Mogo style
> playouts above certain strength, without using larger patterns
> and next
> move choice based on probability distribution. Currently,
> playing out
> a simple joseki in a sensible way in simulations will just
> never happen.
> This is a bit frustrating since all my attempts at
> implementing probdist-based playouts have failed so far, but I
> I will just have to try again...
> To implement softmax, you can refer to my thesis where I have
> described the framework of the move generator for the playout.
> Detecting forbidden moves and replacing useless moves by better
> alternatives are very useful. There you can gain a lot by applying
> much Go-knowledge. Two good candidate algorithms for training the
> feature weights are MM and SB(Simulation Balancing). I tried hard but
> failed to measure any improvement from SB gammas (trained on 9x9) on
> 19x19. You can use CLOP to tune the MM gammas which are far from
> optimal according to our experience.
> Also, my regression test of seki and L&D that pachi has participated
> could be helpful to improve program's tactical strength. In my
> opinion, that is the most crucial factor to reach high-dan level.
> Computer-go mailing list
> Computer-go at dvandva.org
More information about the Computer-go