Step 3: - to build brains

Discussion in 'Intelligence & Machines' started by hlreed, Jan 4, 2003.

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

    Messages:
    245
    We have data streams and nodes to do arithmetic on them. This becomes an algebra and leads to a language to build brains.

    A node is a function of two data stream variables input and one data steam variable output. Internally the nodes all have the variables I for input, R for other input and O for output.

    The code is computer like with differences. ; delimits comments. The main difference is that all nodes/ equations are active at the same time. There are no sequential controls here.

    O = I - R ; is a CNode (tm)
    O = I + R ; is a PNode (tm)
    O = I o R ; is an ONode (tm) O = max(I,R) is arithmetic or
    O = I a R ; is an ANode (tm) O = min(I,R) is arithmetic and
    (O = I * R ; is MPY) if needed and O = I div R is integer divide.
    These nodes all make HalTrees of any size and can match any neuron.

    There are obvious conditional nodes.
    O = I is R ; O = I if I = R else O = 0. There are {is, gt, lt}
    O = I waseq R ; If I=R then m=I and O = I else O = m
    {waseq, wasgt, waslt}
    There are more nodes, such as conditional switches, but, like Boolean algebra, everything can be done with the first four.

    ; Let SL and SR be sensors and ML and MR be motors such that:
    ; If ML > MR turn right.
    ; the code is for seeking whatever the sensor sense.

    MR = SL - SR
    ML = n(MR) ;n is a negate function
    ;end of code

    The truth table for this is:
    SL SR ML MR comments
    0 0 0 0 stop
    0 1 -1 1 turn left
    1 0 1 -1 turn right
    1 1 0 0 stop

    This is all for now. If you like I will send you Hal Algebra for free. Just ask.

    Harold
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Cthulhu_X Registered Member

    Messages:
    25
    HALgebra....
     
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. hlreed Registered Senior Member

    Messages:
    245
    OK.
    Make me another word. Machines move and manipulate force and action. Add a brain and you have a machine that moves data and produces action. It becomes sort of animal and machine so I call them animachines. I don't know if that is the best word for them.
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Rick Valued Senior Member

    Messages:
    3,336
    Like it?
    This is amazing Stuff!!

    Finally we have someone who talks Business.

    Okay now for some doubts...

    Please Register or Log in to view the hidden image!


    • What you are saying is that Node is a Function or are you trying to model a Brain as a transform Function?(I mean Function Modelling?)
    • Also a Haltree is made up of Nodes,Now is Haltree a Neuron?

    Very interesting indeed.

    bye!
     
  8. hlreed Registered Senior Member

    Messages:
    245
    Thanks,

    HalTrees match neurons exactly by number of inputs. I do not have molecular arithmetic yet, so I have modeled neurons as functions, as input, output machines.
    You can use this to build or analyse. Suppose you have a list of motors and sensors for an animal. Divide sensor by motors and you get functional complexity. (Each motor is a function.) That tells you how many inputs you need. You buld that up from HalNodes, starting at the motor, a node, add two nodes to that, add four nodes to those, and keep adding until you get to the number of inputs.
    There are some 20 functions that can be mixed in any manner in the tree.
    Mainly you need either to sum or to choose. Use PNodes to sum, use ONodes to choose. Use CNode to maintain direction of data.
    That is how you buld up a complex machine.

    Check my web site for more info.
     
Thread Status:
Not open for further replies.

Share This Page