In brief, the article describes a graphic of different people’s mental activity and describes in detail how it all works. Not many are able to read the article to the end, even I don’t understand much of it.
If you feel bad about not being superficially familiar with the structure of the brain, I must reassure you that no one understands it completely, even those who claim otherwise. In general, neural networks are pretty well known about, what they do, about their inputs and outputs. The black box in its purest form. There’s a circle of experts who are confident to talk about which part of the brain is responsible for what, except that it is all beautiful guesses and hypotheses. Every year there are new experiments, research and discoveries that contradict previous assumptions.
Let's look at the basics. Neural networks consist of neurons connected by nerve fibers. Via fiber from neuron to neuron impulses are transmitted. This happens when a drop of rain falls on the skin, or when you exchange glances with a hottie, perhaps somewhere on the street you heard your favorite song or even had a nightmare. Thoughts and actions in our body are implemented as transmission of electrical impulses. This process resembles a stadium full of fans when they all rise from their seats and forma “wave”. A neuron is like a weird alien - on one side of it's body there extend many thin threads, dendrites, and on the other side, one thick thread, the axon. Dendrites receive signals and the axon is able to transmit them. It makes sense that the axon of each neuron is attached to the dendrites of another. This fixture is called a synapse. The synapse is not so straightforward as it may seem at first glance, it is able to adjust the strength of the pulse, like the volume knob on the speakers of your computer.
At my request, my colleagues drew me a formula to illustrate a neuron with three dendrites, and the synapses on these dendrites have weights w1, w2 and w3. Let our synapses receive impulses forces x1, x2 and x3 respectively. Then after passing the synapses and dendrites, the neuron receives impulses w1x1, w2x2 and w3x3. The neuron converts the obtained total momentum x=w1x1 w2x2 w3x3 in accordance with a certain transfer function f(x). The power of the output pulse is y=f(x)=f(w1x1 w2x2 w3x3). Thus, the neuron is completely described by its weights wk and the transmission function f(x). After obtaining the set of numbers (vector) xk as inputs, the neuron generates a number y at the output.