a neural network capable of influencing it's own biases, simulating learning to a certain extent. currently pretty useless, but it an interesting proof of concept either way.
brain
learning
machine
network
neural
electricity
electornics
circuit
neuron
computer
Comments
-
yes :)
-
ok so its fine that i blew it up with my azure :D
-
no? it is just "encouraged" not to pick it again as the punishement causess it to alter it's biases.
-
does that mean it feels pain
-
it gives outputs that cause it to either receice a reward or punishment that influences what outputs it makes in the future
-
what do it do tho? 1+ anyways
-
its a refence to the statue called the thinker
-
the thinker??? ace attorney reference???
-
It can and I encourage people to experiment with applicaitions such as that, but I will not be using this model anymore as it is outdated and I have already made a much better one using the LRSLM "lobes" that learns to pick one output over another and does so with suprising accuracy.
-
And that is why I propose this as a variant of the problem that this "neural network" could solve. If it can't, then it can't be called a neural network.