How to determine which neurons to connect between layers in an artificial neural network? -
say first input layer have 10 input nodes/neurons. hidden layer has 10 neurons well. third , final layer 1 output neuron.
how connect layers? there technique determining best way or connect each input neuron each of hidden layer neurons total of 100 edges between 2 layers?
this basic question haven't seen many concrete examples. examples did find either seem connect neurons or connections seem randomized.
if anything, based on intuition , empirical results. i've seen people use recursive neural networks.
with feedforward neural network, makes sense connect neurons layer n
neurons in layer n+1
.
here example latest usage (to demonstrate enormous amount of edges):
Comments
Post a Comment