Neural network in pictures
Infographics: how neural networks work
A neural network is called software capable of learning. For example, by receiving a training sample from a set of recognizable images (for example, a set of people's faces), she is then able to find among them a given face of a particular person. Similarly, it can be taught to "draw", "generate" music and texts — and not only.
In the most general form, neural networks resemble the structure of our brain: there is a processing "neuron", there are "synapses" (connections between neurons), there are transmitted signals. And yet it's only about approximation. How exactly do such networks learn what they should do?
In order for neural networks to learn how to generate an image or something else, they are "trained". To do this, the input information is compared – a phrase (for example, a person asks to draw a picture "about an elephant") – with the output information (a picture "with an elephant").
Portal "Eternal youth" http://vechnayamolodost.ru