GMU:DeepVisualMemories: Difference between revisions

Line 14:

Am Ende des Kurses steht eine Live-Aufführung des entstandenen Werkes.

____

Using a specially designed online learning environment, we first explore the worlds of the cloud in which machine learning (AI) is at home and try to understand its pictorial understanding. This technical excursion is followed by a critical analysis of the topic using selected theoretical writings. At the same time, we will use the process known as Pix2Pix to train neural networks with our own visual condensates, which then crystallize into their structures in the form of learning processes. The resulting condensed crystal can then be used to transfer any visual content into its internal monadic structure.

Together we will develop a project in which we train an AI system, a so-called GAN, with visual material. This system is then able to transfer a visual (procedurally created and music responsive) input into the stylistics of the trained material. The end result will be a performable, interactive live visual system that combines creative expression with the modern possibilities of AI. It will become clear that such a technology can never replace artists, but like any other Toll, it opens up new possibilities for creative processes. In addition, the inner workings of the systems will be revealed and exemplified.

@@ Line 14: / Line 14: @@
 Am Ende des Kurses steht eine Live-Aufführung des entstandenen Werkes.
 ____
+Using a specially designed online learning environment, we first explore the worlds of the cloud in which machine learning (AI) is at home and try to understand its pictorial understanding. This technical excursion is followed by a critical analysis of the topic using selected theoretical writings. At the same time, we will use the process known as Pix2Pix to train neural networks with our own visual condensates, which then crystallize into their structures in the form of learning processes. The resulting condensed crystal can then be used to transfer any visual content into its internal monadic structure.
 Together we will develop a project in which we train an AI system, a so-called GAN, with visual material. This system is then able to transfer a visual (procedurally created and music responsive) input into the stylistics of the trained material. The end result will be a performable, interactive live visual system that combines creative expression with the modern possibilities of AI. It will become clear that such a technology can never replace artists, but like any other Toll, it opens up new possibilities for creative processes. In addition, the inner workings of the systems will be revealed and exemplified.