4 ms·
Nobody is attempting to have one-to-one correspondence between neurons and artificial "neurons", the fact that a single biological does much more doesn't imply
by PeterisP 3y ago
Nobody is attempting to have one-to-one correspondence between neurons and artificial "neurons", the fact that a single biological does much more doesn't imply some limitation or incompleteness (as long as the same computations can be implemented simply by having more of them, and as far as we understand, that seems to be the case) - the choice is primary because due to how our hardware parallelization works, we'd prefer to implement the exact same behavior with 1000 structurally identical simple "neurons" rather than have a single more complex "emulated neuron" that requires more complicated logic that can't be straightforwardly reduced to massive matrix multiplication.
- stcredzero 3y agothat requires more complicated logic that can't be straightforwardly reduced to massive matrix multiplication What are the rationale for thinking it doesn't matter? Or is it the case that researchers think a larger number of simpler neurons could be equivalent to a smaller number of "complex emulated neurons?"
- PeterisP 3y agoYes, computation as such is sufficiently universal that it's very reasonable to assume that a larger number of simpler neurons would be equivalent to a smaller number of "complex emulated neurons".