Wednesday, August 13, 2014
astro picture for the day
Image Credit & Copyright: ESA/DLR/FU Berlin (G. Neukum)
some thought for the day extras. these are notes from conversations elsewhere,
Without the feedback loop to the real world that comes from being embodied there will be no AI. Ungrounded symbol-shuffling attempts at AI such as Cyc have always failed. Simulations are and will always be crude due to scaling laws. Only the real world has the complexity, coherence and detail needed. Certainly a lot of overhead can be eliminated in the autonomic nervous system, but there is a good deal needed to operate a body in the real world.
These researchers are operating with an oversimplified and wrong idea of how neurons operate, which is even further simplified in the hardware design to the point of being a useless caricature of real brains. They haven't even begun to model the sonic and mechanical-resonance aspects of neural transmission and plasticity. They still use the disproven Hodgkin-Huxley model (rather than Heimburg's soliton model) as their base, and virtually ignore the necessity of low-speed impulses to neural function, substituting high-speed communications that do not fulfill timing and phase coordination functions. Cytoskeletal dynamics and their effects on morphology and connectivity are entirely ignored, the synapse connections being put in by hand and with crude rules-of-thumb.
Maybe some way can be found to make it work to some extent, but they don't even seem to know what they're sweeping under the rug. They have an infinitesimal chance of success at producing a real AI no matter how many billions of these poor approximations to neurons they manage to pile up." - MUltan( that's what he calls himself)
for which I replied,
"I've always agreed that life is a stable non-equilibrium structure as Ilya Prigogine argues.
I'm surprised the A.I. guys don't get into Ilya Prigogine's stable structures ideas; seems to me the idea that life is connected to the environment through numerous feedback loops fits in with Marvin Minsky's 'Society of Mind' idea. The society of mind . . . ideas mixing and matching(some would say competing) would come from these numerous feedback loops;" - Flashgordon!