[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [computer-go] Designing faster, better influence functions



For the task of estimating potential territory my experiments showed that linear influence functions are not worse than, e.g., Zobrist's method. At least for this task, the improvement by using a non-linear influence function is relatively small. I would not be surprised if for many purposes the quality/speed trade-off would favor a linear mapping.

Erik


Gunnar Farnebäck wrote:
Ok, that's where we differ. I don't consider an influence function
which allows influence to flow through other (living) stones to be
interesting, but as usual it depends on what you want to do with the
influence.
Tom Cooper wrote:
> However, I share your skepticism about the usefulness of a linear influence
> function.

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