What are neural networks?

Neural networks are computational models. They consist of "artificial neurons". In principle, neurons are the smallest units for information processing, which work in parallel. These networks have their name because their structure and function is similar to the structure of biological neural networks - for example, the human brain.

An artificial neural network is a model of a system of artificial neurons (nodes) and their connections. The arrangement of the nodes resembles individual layers. If the network consists of many layers, it is called deep learning.

Artificial neural networks mimic the way the human brain functions - they can recognise patterns, are capable of learning and can generalise what has been learned. They learn from huge amounts of data and are used today, for example, in forecasting, data classification or facial recognition.