Step 1 nature ->sensors->brain->motors->nature

Discussion in 'Intelligence & Machines' started by hlreed, Dec 30, 2002.

Thread Status:
Not open for further replies.
  1. hlreed Registered Senior Member

    Messages:
    245
    At the beginning, look at any living thing. It is an entity with self in an environment. Call the environment nature, then the data flow of any active body is:
    nature -> sensors -> brain -> motors -> nature
    The self is sensors, brain and motors. From the data flow note that brain is singular and is always between sensors and motors.
    A sensors reads an aspect of nature and abstracts it to a number.
    Brain reads all the sensors, makes a function for each motor and writes data to each motor. Motors read the brain and convert the numbers into an action that pushes on nature.
    A neuron is a brain function. The synapses are virtual motor inputs and the axon is a virtual sensor output. Let the axon be A and the synapses Sx where x is a number. Then
    A = N(S1,S2...Sn) is a function N.
    Brains are a list of sensors of neural functions. Brains are of any size, from 0 to n.
    Now we can compute brain functional complexity. Count the sensors and motors. Divide sensors by motors and you have the number of neural functions.
    This is step one in building brains. In step two we will make some neural functions and neurons.
    Harold
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Cris In search of Immortality Valued Senior Member

    Messages:
    9,199
    OK. Don't let me stop you. Carry on.

    I'd be interested in seeing a computer model of whatever you can design.
     
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. hlreed Registered Senior Member

    Messages:
    245
    Thanks Cris,
    We have neural functions now. Next step is to see what the functions are.
    Take C. elegans, the only worm with a parts list. It has about 14 motor functions. That is 84 sensors if we use 6. 84/14 = 6 which is 6 sensors per function. The worm has about 210 neurons that are not sensor or motor. This means 15 neurons per motor function.
    I have evened up the numbers so all functions are alike.
    An = Nn(s1,s2,s3,s4,s5,s6) where sn are sensors or inputs.
    When you make a fully connected tree with two input nodes you can calculate the number of nodes (neurons) required to fill it.
    A six input function or tree requires 15 nodes to fill it. That is because the number of nodes goes down by one in each layer.
    6 + 5 + 4 + 3 + 2 + 1 = 15 is the tree in layers.
    Because it is a full tree we can write the function.
    An = s1 o 5*s2 o 10*s3 o 10*s4 o 5*s5 o s6
    o is any operator. (+,-,o,a,is,gt,lt,waseq,waslt,wasgt are some tree functions.
    This is an interlude. Step two will make some functions and function machines.
    Harold
     
    Last edited: Dec 31, 2002
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Cris In search of Immortality Valued Senior Member

    Messages:
    9,199
    Why C.elegans, there are many types of nematode?

    Who are you and what are your goals? Is this a hobby or an occupation?

    Has anyone fully emulated a complete nematode yet, to your knowledge? What are the obstacles?

    Take care.

    PS, if you can then some more white space in your text would make your posts easier to read. Thanks.
     
  8. hlreed Registered Senior Member

    Messages:
    245
    I got introduced, seduced by Marvin Minsky in 1958 when he was selling AI.

    I gave it up in 1975 or thereabouts when I realized nothing could come of AI because intellegence is not a thing and a computer has no brain and even if you made a brain it could not get out of the computer. At that time I was trying to make spoken language into machinery with AI and it failed. At that time I also became old enough not to listen to authority or experts. (Now I am one.)

    So I started with the obvious, looking for the least brain. I found that in 1994 and wrote a book about it. Book title is
    Brains for Machines/Machines for Brains
    It was not very good and noone understood what I was doing. One reviewer called it "Wierd". Another thought the first part was good, while I thought it was filler, and the last part was nothing, except my message.

    The world is getting very close to this now. The Japanese will discover it soon if not already. The major change one must make is to make motor functions from yin - yang weights rather than by if - then rules.

    I have a company, Hal Brain Design that offers brain kits and free tutorials and so on, but there has not been much interest.

    Rodney Brooks said I should make my nodes programmable, while I want to move programming to the machine I am building.

    This is serious stuff. This is me.

    Harold L. Reed
    Hal Brain Design
    2012 Melody Lane
    Richmond, TX 77469
    281 232-4466
    hlreed@halbrain.com
    http//www.halbrain.com
     
  9. Cris In search of Immortality Valued Senior Member

    Messages:
    9,199
    Harold,

    Are you aware of the WBE (Whole Brain Emulation) project and the current research being undertaken by MURG (Mind Uploading Research Group)?

    Here is a link to their C.elegans nemotode upload project.

    http://www.minduploading.org/research.html

    Is this of use to you?

    My primary interest in this arena is Minduploading but I don't believe I have the skill set to make significant research contributions, so instead I content myself with monitoring their progress.

    While I see the development of robotics as vital to our future it is the emulation of the human brain that is of much greater importance. While this is doubtless decades away I would hope that we could make enormous progress in the meantime by completely emulating the brains of simpler organisms. The nematode is the obvious start.

    I can’t establish from your posts where you are in this endeavor. Do you have an accurate emulation of the nematode brain?
     
  10. hlreed Registered Senior Member

    Messages:
    245
    Cris,
    Hal algebra is a tool to both analyse and synthesise brains. It is free from my website.

    C. elegans is an example of how one analyses a small brain. I have no interest in its brain, I just happen to know the numbers.

    I am currently working on sensor to percept to concept. I know how to build and compare concepts. I do not know how to make sensor data into concepts.

    In vision, one can define an object as a self similar patch. There are lots of ways to do this, but I have not been able to build large objects yet.

    I have the code for Big Al, a human sized animachine. Al knows how to sit, stand, walk and so on, but does not know what to sit on or where to walk.

    I hope that once enough people know the algebra that we can work together on stuff like this.

    See step 2.

    Harold
     
Thread Status:
Not open for further replies.

Share This Page