I’m not a biologist but I study computer science and have a basic understanding of neural networks. Neural networks are inspired in actual neurons and glia.
A basic neural network is made from a layer of input neurons, various layers of processing neurons and a layer of output neurons. Neurons are connected to each other by paths that has their own value and neurons themselves have a logical component. For example: if the addition of all the paths that are connected to me is over x value send signal. When we train neural networks we are actually changing the values of the paths between neurons. It is agreed that the information contained in a neural network is actually on the values (weight) of those paths, however, it is basically impossible to understand that information.
Basic neural networks are usually trained in problems of classification or regression because they do a really good job.
Latest Answers