The basis of deep learning are so-called neural networks which imitate the human brain. The neural networks enable artificial learning by constantly being re-linked with and amongst each other as part of the learning process.
Unlike machine learning, however, people do not intervene in the deep learning process. The human being only supplies the system with all the necessary information and ensures that the processes are documented. The analysis, the derivation of decisions or forecasting, are independently executed by the system. Accordingly, humans have no influence on the results, nor is the decision-making process comprehensible. The system is subject to an ongoing optimization, as the decisions taken are constantly questioned and, if necessary, the pattern underlying the decision is adjusted.