Neural Net: A New Kind of Processor (HELP)
-
@tothemyers (View Post)
No, because the Wiki does not lie. I've been testing it. The neutrons go through the glass and turn into sparks when they go to the PSCN/ NSCN. keep trying...
Try something like this:297593View Save 297593
(don't vote this save by me up, as the basis of this is not by me) -
Collide photons with breakable metal for random outputs.
If you want 4 random outputs.
1. collide a photon with broken metal (50% will pass through, 50% will reflect)
2. collide the reflected photons and the photons that passed through with two more pieces of broken metal (this will create two more situations where the photon will have a 50% chance of either passing through or reflecting)
3. Use semiconductors to convert the photons to sparks for the outputs.
You know have 4 outputs each with a 25% chance of occuring.
I guess you would also need to convert the spark back into a photon for the next neuron. Or you can just bypass converting it to a spark and use photons outright for sending information. -
-
-
@vanquish349 (View Post)
tothemyers only thinks it doesn't because the chance that it does is not very high. It takes a fair bit of neut in that environment to form a photon that will spark. -
-
@The-Con (View Post)
Thanks for that excellent neuron. I will be using a (very) slightly modified copy for the creation of the neural net. You, along with the others will receive partial credit when the final copy of the neural net is released.
I would also like to apologize for being so naive when it came to neutrons turning into sparks on semiconductors, in calling the wiki (like the cake :) a lie. My testing conditions were not proper, so I am sorry for doubting what all of you + the wiki said.
Now that the neuron stage is done, it gets to the tricky part: Designing a network that reacts to info in a seemingly coordinated way. I am, in all essentials, trying to mimic the human brain. -
[quote=mniip]I think TPT screen will be TOO small.../quote]
If gnat has a small brain, but still responds intelligently to info. If a gnat, with it's pinprick of a brain, can do it, I can too. The problem so far is reducing the output from neurons from several to one. It has to actually have coordinated processing, not just look like a nuclear reaction.
