oThe main difference
between a SOM and a competitive network is that in the SOM numerous
neurons are embedded in a large mesh.
oThe neurons
are divided up into neighborhoods (many local maxnets) and go
though cycles of inhibition and excitation.
oThe mesh
determines the excitatory and inhibitory relationships.
oAfter
some number of cycles the neurons congregate in the weight
space in such a way that they map out areas of high
probability.
o
Application
of Neural Networks to Phylogenetics - SOTA