Training data synthesis is a term used in the fields of artificial intelligence, big data and smart data as well as Industry and Factory 4.0. It describes the artificial creation of data that is used to train AI models. Instead of relying exclusively on real data that is often difficult to access or sensitive, new, artificial training data is generated with the help of certain algorithms.
The aim of training data synthesis is to make the development of artificial intelligence easier and safer. Real data is often too expensive, difficult to obtain or contains personal information that is protected by data protection laws. These problems can be avoided with synthesised data. At the same time, it is also possible to simulate rare or dangerous situations that rarely occur in reality.
An illustrative example: A company wants to develop an AI for quality control in a car factory. Instead of collecting thousands of images of real defective car parts, the company uses training data synthesis to generate artificial images that show different types of defects. This allows the AI to learn faster and more efficiently without the need for real defects to occur in the factory.
Deutsch 
English (UK)