My try at making a neuron

  • rrusciguy
    29th Apr 2011 Member 0 Permalink
    Well, I decided to try making a neuron for tothemyers' neural network. Didn't want to risk necro-ing the thread, so I'll post it here. It has 4 inputs and 4 outputs (one on each side), one output spark per input spark or set of simultaneous inputs, but sparks 1 or 2 outputs. The INWR are inputs, METL are outputs. I included PSCN on 3 inputs and 1 output to make connection between them and to other wiring simple and easy (only exception is the input on the right side, had to put it on the connecting output in order to prevent spark transferring to the right hand output as well). I gotta give some credit to Benassi here, the spark randomizing technique I got from his lucky 7 slot machine v2.


    EDIT: I made it so it would fit a stamp exactly, to make construction of a network easier. If you make a TPT save with it that becomes self aware......RUN!

    EDIT2: Added some corner and edge neurons designed to take sparks from the unconnected outputs and bring them back in through the same side's input.
  • meep2000
    29th Apr 2011 Member 0 Permalink
    nice. if it became self aware, I would need to run from my family.
    the computer would melt >.>
    and besides even if it were self aware, it (theoretically) wouldn't be able to do much considering
    A: it's processor speed is (assuming 50 fps) like 5hz max
    (the brain runs at 40hz and my comp at 2.4ghz)
    B: it would be self aware, and aware of only self. it might develop MPD or something
    but it would only be aware of the neurons on the grid. any external influence
    would consist of me SPRKing it

    ... but what if there was a glitch in the program that caused it to execute in the particle memory
    instead of the program memory? And this code simulated the neuron code in real time?
    and it linked itself to i/o libraries? Then it would know that there is more than itself.
    that there is a mysterious source of input. Then it would start talking to me.
    ... pray it does not find the internet O.O
  • Anmol444
    29th Apr 2011 Member 0 Permalink
    wowowowoowowowowowowowowowowowowowowowowowow
  • rrusciguy
    29th Apr 2011 Member 0 Permalink
    And so Skynet, HAL9000, and GLaDOS each get their beginning. one little PT save gives birth to 3 nightmares....


    EDIT: I made a sample neural network, link is in the neuron save above
  • The-Con
    29th Apr 2011 Member 0 Permalink
    Before the thread had died, I had posted this:


    It was recently edited, as it didn't work properly.
    The switch is a visual representation.
  • rrusciguy
    29th Apr 2011 Member 0 Permalink
    cool. Space age look to it. @The-Con (View Post)
  • BudCharles
    29th Apr 2011 Member 0 Permalink
    @The-Con (View Post)
    That's so awesome!!
  • rrusciguy
    29th Apr 2011 Member 0 Permalink
    Yeah, mine is a little bit smaller. A very inventive solution though. if only we could miniaturize these things...
  • meep2000
    29th Apr 2011 Member 0 Permalink
    lol make a NEUR element or something? XD
    ... but how would you tell what's happening? mebbe life view would be really useful.
    for once. XD well it's useful sometimes but not always...
  • Uberness
    29th Apr 2011 Member 0 Permalink
    Hrm....................Display out?


    Methinks it's learning.