Re: [computer-go] TD(lambda), Neural Networks and evaluation functions

[Peter Drake <drake@xxxxxxxxxx>]

It did improve -- the trained network could easily beat the untrained 
network.  However, it was still so lousy that my students (who were 
around 25 kyu) could trounce it easily.


On Tuesday, September 16, 2003, at 03:09  PM, Imran Ghory wrote:

> On Tue, 16 Sep 2003, Peter Drake wrote:
>
> > Yup, we did this over the past summer, with not too much success.  I'm
> > now convinced that it is necessary to give the program more structure,
> > e.g., some way of modeling chains/blocks, connections, eyes, and so 
> > on.
> >   This is too much for a generic feed-forward neural net to derive by
> > itself in any reasonable amount of time.
> Was it improving at all (i.e was the predicition made say 5 moves from 
> the
> end becoming more accurate) or was it just a case of the prediction 
> being
> little more than random ?
>
> Imran
> --
> http://bits.bris.ac.uk/imran
>
> _______________________________________________
> computer-go mailing list
> computer-go@xxxxxxxxxxxxxxxxx
> http://computer-go.org/mailman/listinfo/computer-go
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/

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