So if a column isnt predicted, all cells become active. My system allocates synapses randomly to the data, so if I allocate all the synapses of the dendrite to one column then itll pretty much always be all cells active.
I thought of a solution, per dendrite I only allocate once at a column no matter how many cells active, and that sorta sounds right.
But the problem of forking still remains in my head.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
So if a column isnt predicted, all cells become active. My system allocates synapses randomly to the data, so if I allocate all the synapses of the dendrite to one column then itll pretty much always be all cells active.
I thought of a solution, per dendrite I only allocate once at a column no matter how many cells active, and that sorta sounds right.
But the problem of forking still remains in my head.
i read another post, and yes this is indeed a problem we must think about. i will finish my implementation and put it here tho.