Deep Flux

Deep Flux

Artist: Judit Eszter Kárpáti

Title: Deep Flux

Attribution: Deep Flux, 2022, Judit Eszter Kárpáti and Esteban de la Torre (EJTECH)

Year: 2022

Materials: Textile, reactive pigment, custom code and electronics, extruded aluminium, Raspberry, Manfred Mohr and Vera Molnar works

Dimensions: Variable Dimensions

Image Statement: Deep Flux signifies the dawn of an era where the digital constrainments of computation, and our interaction with pixel displays is replaced by dynamic, desiring metamaterials in constant flux. Using Generative Adversarial networks (GAN) an AI is trained to perpetually paint new paintings on a canvas treated with reversible dynamic pigment. This enables the neural network to reach out into the physical world and create short lived traces and compositions in constant flux. Training datasets are based on curated works from pioneers of computer art Vera Molnar and Manfred Mohr. Confronted with this process, the viewers' tendency to impose meaningful interpretation and narrative to the generated art piece questions whether art lies in the creative emancipation of the AI, the black pigment potentially containing all colors, or the intercognivite experience of the installation. Exhibiting together with Vera Molnar's Les Métamorphoses d'Albrecht and Hommage á Dürer, as well as Manfred Mohr's Golden Nica winning work P-412-C. These as well are among the works used in the training dataset for the AI.