markh@csd4.csd.uwm.edu (Mark William Hopkins) (12/01/90)
In article <3415@bruce.cs.monash.OZ.AU> frank@bruce.cs.monash.OZ.AU (Frank Breen) writes: >In one book by Richard Dawkins (the Selfish Gene I think) He >estimates roughly how many bits per second of input the brain >is processing. I can't remember the details, but it seemed >surprisingly low, of the order of real time video. > >He worked it out from the definition of our vision, hearing etc. > >Do people (who've read it) think this means anything or what. > >-- >Frank Breen "I am a warrior in the landscape of my mind" >Monash Uni -unkown? It's a well-known result that visual processing takes place in the brain too fast and signals in the brain travel too slow for visual signals to pass through more than (say) 5 to 10 layers of neurons before having been fully processed. It means that there cannot be any complex iterations, searches, or other similar forms of computation going on during recognition: i.e. no sequential symbolic processing. Did you ever play around with a prism? It effectively computers the Fourier Transform of an input signal by shuffling light quanta among a lattice of particles (or something like that): all in parallel and all in an extremely short time ... but yet uses no symbolic algorithm to accomplish its result. There's an analogy for you...
minsky@media-lab.MEDIA.MIT.EDU (Marvin Minsky) (12/01/90)
There are some thoughtfully worked out estimates in Hans Moravec's book, "Mind Children", Harvard Univ. Press, 1988. I can't find my copy at the moment. I recall that he predicts that given his brain-power estimates, against his predicted growth in computer memory, speed, size, and costs, that a brain-equivalent computer will be of desktop workstation size and cost in from 30 to 40 years. Please don't argue about this until someone else (or me) copies out Moravec's actual prediction data, because I don;t have the details at hand. Most of those 100 billion brain cells are very small ones, like granule cells, which have only a handful of synapses, and probably only a small proportion of them have the order of 10000 synapses. But many have 1000. Adding up these numbers is an interesting but somewhat degenerate activity. Kandel and Schwartz point out that there are between 1000 and 10000 different TYPES of neurons in the mammalian brain. Makes one think, doesn't it.
wcalvin@milton.u.washington.edu (William Calvin) (12/02/90)
My standard refrain about the numbers game (for neurons in the
brain), back when I wrote THE THROWING MADONNA, was "that of those
famous 10 billion neurons comprising the brain, 100 billion of them
are located in just one section of the brain (the granule cell
neurons of the cerebellum)." The number 10 billion was the 1968
estimate of the numbers of neurons in cerebral cortex (which is not
exactly all of the brain) of one hemisphere (so 20 billion in both).
More recent estimates more than doubled that. But estimating whole
brain is still a very dicey business, even for order-of-magnitude.
The number of synapses per neuron is somewhere between 1E3
and 1E5, and depends on neuron size.
William H. Calvin wcalvin@u.washington.educam@aipna.ed.ac.uk (Chris Malcolm) (12/05/90)
In article <4208@media-lab.MEDIA.MIT.EDU> minsky@media-lab.media.mit.edu (Marvin Minsky) writes: >Kandel and Schwartz point out that there are between 1000 >and 10000 different TYPES of neurons in the mammalian brain. Makes >one think, doesn't it. Well, that's the current working hypothesis :-) -- Chris Malcolm cam@uk.ac.ed.aipna +44 31 667 1011 x2550 Department of Artificial Intelligence, Edinburgh University 5 Forrest Hill, Edinburgh, EH1 2QL, UK DoD #205