With the biological knowledge of the way in which we think that dendritic connections create associations/memory, particularly the proximal and distal sensitivities, if you were starting a code base from scratch would you write a solution that is the same as the current implementation ? The logic of an arbitrary fixed size SDR limit is very much a computational work around and a breaking / capping of biology plausibility for example.
The reason for the question is to ask about variations on thoughts about how the process works biologically and the way in which the code emulates or approximates the process we think is happening inside the blob on our necks. How do you think about the process as a pattern and does this pattern match the current approach ? Which bits of it just feel odd or not quite right ?
My opening thought would be to change the perspective / calculation to a more dendritic centric type of approach rather than a neuron centric approach. Thoughts ?