Re: f(x) and NN

[Mousheng Xu <moushengxu@xxxxxxxxxxxxxxxx>]

At 09:18 PM 5/6/99 -0700, Weimin Xiao wrote:
>Hi, there,
>
>I would think P.J. Leonard's position = f(situation) function goes a little
>bit too far. A simpler expression can be
>    value = f(board)
>which is what we know as an 'influence function'. Current influence

	Should it be only an "influence function"?

>Why existing NN approaches does not work well? My guess is people either
>used
>too few hidden nodes or too many layers. A three-layer BP NN without last

	Maybe "too few hidden nodes". I bet most if not all of them use 3-layer. 
	Now, could you estimate how many hidden nodes you need? 

>How many hidden nodes are needed? Roughly, one pattern one hidden node if
>the
>neural network has not been 'generalized' - if we guess it will be
>generalized, then I do not know how few nodes will be used.

	One pattern one node is hopeless, as you stated. As least you don't have
disk space for all the possible patterns. I strongly agree that some sort
of generalization is needed for NN. But how? Any discussion on this subject?
	Thanks.

-- Mousheng Xu