On Sun, Apr 1, 2018 at 12:29 AM, Lawrence Crowell <[email protected]> wrote: > On Saturday, March 31, 2018 at 2:32:06 PM UTC-6, telmo_menezes wrote: >> >> On Sat, Mar 31, 2018 at 10:17 PM, Lawrence Crowell >> <[email protected]> wrote: >> > You would have to replicate then not only the dynamics of neurons, but >> > every >> > biomolecule in the neurons, and don't forget about the oligoastrocytes >> > and >> > other glial cells. Many enzymes for instance to multi-state systems, say >> > in >> > a simple case where a single amino acid residue of phosphorylated or >> > unphosphorylated, and in effect are binary switching units. To then make >> > this work you now need to have the brain states mapped out down to the >> > molecular level, and further to have their combinatorial relationships >> > mapped. Biomolecules also behave in water, so you have to model all the >> > water molecules. Given the brain has around 10^{25} or a few moles of >> > molecules the number of possible combinations might be on the order of >> > 10^{10^{25}} this is a daunting task. Also your computer has to >> > accurately >> > encode the dynamics of molecules -- down to the quantum mechanics of >> > their >> > bonds. >> > >> > This is another way of saying that biological systems, even that of a >> > basic >> > prokaryote, are beyond our current abilities to simulate. You can't just >> > hand wave away the enormous problems with just simulating a bacillus, >> > let >> > alone something like the brain. Now of course one can do some >> > simulations to >> > learn about the brain in a model system, but this is far from mapping a >> > brain and its conscious state into a computer. >> >> Well maybe, but this is just you guessing. >> Nobody knows the necessary level of detail. >> >> Telmo. > > > Take LSD or psilocybin mushrooms and what enters the brain are chemical > compounds that interact with neural ligand gates. The effect is a change in > the perception of consciousness. Then if we load coarse grained brain states > into a computer that ignores lots of fine grained detail, will that result > in something different? Hell yeah! The idea one could set up a computer > neural network, upload some data file from a brain scan and that this would > be a completely conscious person is frankly absurd.
The molecules of LSD, psilocybin, etc have specific binding affinities to various neuroreceptors. Ok. This is a very important point and I completely sympathize with you bringing it up. Current artificial neural network models are extreme simplifications. We could say that they model a brain that only uses Glutamate (excitatory signals) and GABA (inhibitory signals). The other neurotransmitters are responsible for a lot of interesting stuff, namely learning. It is telling that contemporary ANNs resort to a blunt, centralized algorithm for that part (backpropagation). But how much information is contained in a molecule of LSD? And how much information is necessary to define a receptor site? Imagine a model similar to what John Holland suggested, where you define these things as strings of letters. Let's say with an alphabet of four letters (a, b, c, d) -- because nature seems to like that. So we could have a drug that is abaccaba and we could have a receptor site that is abbccacb. Then use edit distance to determine their affinity. How many letters would we need to model something with the complexity of the human brain? Not a lot I bet. My point is: there is no reason to assume that we have to go into extreme detail, such as molecular interactions with water or quantum states. Maybe we do, but the stuff you allude to could still be way simpler than that from an information theory perspective. Telmo. > LC > > -- > You received this message because you are subscribed to the Google Groups > "Everything List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/everything-list. > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/everything-list. For more options, visit https://groups.google.com/d/optout.
