ABSTRACT
The artificial neural networks are definitely the most popular AI technique currently with the number of potential application limited only with our creativity. The chapter starts with some key information about the human brain, neurons and the structures they form. Then the Reader learns that any form of perception (sound, image, movie) representable in a computer is actually a set of 0’s and 1’s, what makes them usable as an artificial neural networks inputs. The chapter also provides a very careful description of some simple artificial structures together with the explanation (also by analogy to our everyday life) of how these structures are taught (by human teacher) and later successfully used for solving complex classification problems. The concept of learning and testing set is clarified in here too. Various topics related to this AI method are explained as well: layers (network topology), weights (and how they are modified), gradient descent, algorithms like backpropagation and simulated annealing. Deep learning is recalled with its main assumptions and example applications.
