A neural network is essentially one of two ways of modeling the mind, or consciousness.
The traditional method, physical symbol systems, was first proposed. It works essentially the way your computer works; it has a specific set of rules; a large random access memory (RAM) and a small CPU that takes informatino from the RAM, processes it, and puts it back. The power of a physical symbol system is that it does one task at a time, but it does this very quickly.
The problem with the PSS is that it follows a strict set of rules. It doesn't really learn. (Actually, the *real* problem with a PSS is that it can't combine top-up and bottom-up information together, so while it can beat the world's greatest chess champion, it has a hard time recognizing a pattern that a child could recognize.
The next model of cognition is the Artifical Neural Network, or Neural Net.
This model is based on the structure of the brain. Essentially, there is no CPU.
Information processing (IP) as a task is instead distributed every IP unit (modeled after neurons).
The unit, like the neuron, is essentially a switch. it's either "active" or "not" and whether or not it's active is a function of the amount of 'stimulation' (and inhibition) it recieves from its connections, as well as a set threshold level.
Thus the 'conclusions' a neural net reaches is actually a function of the activation patterns of the entire neural net.
A neural net can does learn if there is an outside program that "guides" the neural net. Let's say our neural net is the computer program. A human could be the one "guiding" this learning.
Essentially, we could start off with a completely arbitrary and randomized set of "weights" for each neuronal connection, and over time, if we "correct" the machine for every desirable outcome and vice versa, the machine essentially learns, over time, by trial and error.
some1 can say me how i can buila ar robot who have neural net and laptop wher he save his mind i'am a biginner but i wanna make robot
karistajaify 8 months ago
....Open Claw.....Insert Neck ;)
sweet setup
PsionNinja 2 years ago
*hat down* respect.
blytqb 3 years ago
what soft did u use ?
vacuum20 3 years ago
Awesome, how complex is the neural network?
minousoft 4 years ago
A neural network is essentially one of two ways of modeling the mind, or consciousness.
The traditional method, physical symbol systems, was first proposed. It works essentially the way your computer works; it has a specific set of rules; a large random access memory (RAM) and a small CPU that takes informatino from the RAM, processes it, and puts it back. The power of a physical symbol system is that it does one task at a time, but it does this very quickly.
Joe22c 3 years ago
The problem with the PSS is that it follows a strict set of rules. It doesn't really learn. (Actually, the *real* problem with a PSS is that it can't combine top-up and bottom-up information together, so while it can beat the world's greatest chess champion, it has a hard time recognizing a pattern that a child could recognize.
The next model of cognition is the Artifical Neural Network, or Neural Net.
This model is based on the structure of the brain. Essentially, there is no CPU.
Joe22c 3 years ago
Information processing (IP) as a task is instead distributed every IP unit (modeled after neurons).
The unit, like the neuron, is essentially a switch. it's either "active" or "not" and whether or not it's active is a function of the amount of 'stimulation' (and inhibition) it recieves from its connections, as well as a set threshold level.
Thus the 'conclusions' a neural net reaches is actually a function of the activation patterns of the entire neural net.
Joe22c 3 years ago
A neural net can does learn if there is an outside program that "guides" the neural net. Let's say our neural net is the computer program. A human could be the one "guiding" this learning.
Essentially, we could start off with a completely arbitrary and randomized set of "weights" for each neuronal connection, and over time, if we "correct" the machine for every desirable outcome and vice versa, the machine essentially learns, over time, by trial and error.
Joe22c 3 years ago