In the exhibition TRAINED SCULPTOR, results from the scholarship Module C: INNOVATIVE ART PROJECTS of the BBK Bundesverband are presented.
In it, artist Nikos Probst explores the influence of artificial intelligence in sculptural form-finding processes. For this purpose, he used a so-called Generative Adversarial Network (GAN). The GAN was trained by him with 3D data of 2000 sculptures from different museums worldwide. After the training, the GAN was able to independently generate new 3D models of sculptures from the data set. This data can in turn be used as a template to create real sculptural objects from.
The exhibition features 3 sculptures that were created by the GAN and then placed in real space using classical sculptural methods, such as bronze casting. The presentation of the objects is complemented by a video installation that documents the working methods of the GAN. In it, sculptures can be seen in continuous succession which are created and output in real time by the GAN.