+MrSpudNugget my idea is similar to that! I want the neural network (or brain) to perform, let's say, the AND operation. To emulate learning, we let the circuit output something random first, and then we tell it if it is correct, too low, or too high. The "brain" is then readjusted to better perform the logical AND operation. This way, we are "teaching" the brain how to do the AND operation!
let me rephrase that... a mechanism is there so that anyone who's using it can say if the response was right or not...
and then anyone who's using it can identify a sort of mechanism like "was this right? yes/no?" so it learns what the best response is to a phrase/ question...
is it possible to start to something that responds to basic phrases?
hmmm... now thats where i don't know where to start...
Layering might help with this.
I think that we'll have to move away from the concept of realistic neurons, which would take up the entire screen if we tried to make even a basic intelligence, and start developing advanced thinking devices that are far more advanced than actual neurons.
@msasterisk Yes, that's what I was thinking. To use PROT to mimic a form of "synaptic plasticity" for first level learning.
Would it be possible to change the behavior of the neural network by, say, using PROT to change the temperature of the ARAY in whatever way is desired? That's the only way I know of to change its temperature without using PROP, HEAT, COOL, console, etc.
I'm not great with automation and tpt life, but I'll sure be looking at all these saves!