What is Deep Learning? We explain term & background

For science fiction fans it has been certain for decades that Artificial Intelligence will one day reach unimaginable heights. Currently, Deep Learning is one of the central methods for progress.

Visionary or risky? In Deep Learning, artificial intelligences are supposed to improve on their own.

The more AI advances, the more justified criticism there is, specifically regarding precautions. With Deep Learning, it becomes clear where the risks might lie.

Deep Learning: the human brain as a template

Deep Learning is a subfield of machine learning, also known as Machine Learning.

• Deep Learning uses neural networks, as they also work in the human brain.
• The method consequently tries to imitate the brain as well as possible, in the hope of thus also being able to better understand and map its functioning.
• As a goal of Deep Learning, the AI should be able to make its own decisions based on the available information. The AI thus gets its own thinking ability, which is a key element of "real intelligence."
• This also includes the meaningful combination of different information. How meaningful the result actually is, the AI can subsequently evaluate itself to draw conclusions for the future.
• In the course of the process, more and more neurons are formed and additional layers are created - from this depth the term "Deep" is derived.

Per Deep Learning, AI systems should become independent and improve further in this way. This is at the same time the important demarcation from Machine Learning, since the human is more involved there, for example in controlling decision-making processes. In Deep Learning, humans take on more of an observer role and provide the framework conditions; the actual development is left to the machine itself, and this makes the topic particularly exciting.