Generative Adversarial Networks (GANs) (Glossary)

Generative Adversarial Networks (GANs) belong to the category of Artificial Intelligence and are also important for Big Data and Digital Transformation. GANs are a special technique within the field of artificial intelligence that allows computers to independently create new content. The special characteristic of GANs is that two artificial intelligences work against each other: one AI tries to generate data that is as realistic as possible – for example, images, music, or even texts – while the other AI tries to distinguish real data from artificial data.

This „game of competition“ leads to generated results becoming ever better and more natural. A simple example: GANs can be used to create realistic photos of people who don't actually exist. This technique is used, for example, to generate photorealistic product images for online shops before the actual product has even been built.

Thanks to Generative Adversarial Networks, processes can be automated, costs saved, and new digital possibilities opened up – from image creation and music production to artificial voices. For companies, this technology offers new opportunities to test innovations faster or develop bespoke solutions.