Re: [computer-go] Re: Weights are not very important was: ANNsas potentially useful in the computer goproblem

[David Fotland <fotland@xxxxxxxxxxxxxxx>]

At 05:38 PM 1/11/2004 +0100, you wrote:


> I think the same should hold in Go. I assume it is much more important to
> recognize groups correctly or the divide the game in the correct subgames,
> than optimizing the weights. I also assume that the flat gradient - changing
> weights has only a marginal impact on performance - is also a serious
> problem for an automatic learner. I assume from my hand-tuning experience,
> that there are also a lot of local optimums.
I think this is true.  I spend much more time in Many Faces adding new go 
knowledge and patterns
than I do tuning weights.  The initial weight values I put in are usually 
good enough.

For computer go there is the additional problem or move pruning and 
sorting, since local
and global searches are highly pruned.  This introduces many more 
weights.  If a
particular class of moves is not included, then no amount of weight tuning can
get good play.

David


> Note: I assume of course, that the weights are already in a more or less
> optimum range. The flat gradient "law" is only true around this optimum.
>
> Best Regards
> Chrilly
>
> _______________________________________________
> computer-go mailing list
> computer-go@xxxxxxxxxxxxxxxxx
> http://computer-go.org/mailman/listinfo/computer-go

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://computer-go.org/mailman/listinfo/computer-go