What is the fundamental difference between these two neural network structures?

Dear All,

Could you please help me understand as to what is the underlying difference in the processing of a neural network that receives for example three separate inputs to give a single output this this:

and a network that takes a single input of size 3 to give one output as in this:

If at all there is any difference then where would each work better than the other? Also, what is the best way of assigning inputs to the first network i.e. the one with several inputs?

Many thanks for all your comments in advance. Shashank

Answer 1

[Greg Heath]

The first topology is probably useful if the three inputs come from three completely separate (e.g., physically separate ) sources. Otherwise, I see no reason for it.

I have been designing neural networks for over 35 years and have never had to use that topology.

Hope this helps.

Greg

[Greg Heath]

Please post the code for the first topology. I would like to make a numerical comparison.

TIA, Greg Greg

[Shashank Arora]

Hi Greg,

Thanks so much for your comment. Could you still please help me to figure out how to pass inputs through the first network?

It can be generated using the custom command - 'network'. The code for this particular topology would be:

net = network(3,3,[1;1;1],[1,1,1;0,0,0;0,0,0],[0,0,0;1,0,0;0,1,0],[0,0,1])

Thanks again, Shashank

[Greg Heath]

input1 = {[1] ;[2]; [3] }

or

input2 = [ {1} ; {2}; {3} ]

should work.

Hope this helps.

Greg