The first GAN image
FAKE Seed 01-000160_01
I'm exploring GAN technology to create my Digital Twin. Some questions I want to explore are:
Will the GAN collage images look like something I would make?
What will it reveal to me about my own patterns when making collages?
Social media has become a pervasive tool that offers unprecedented levels of communication and information, but can be detrimental both psychologically and creatively. I will post these GAN generated collage images on social media along with my hand-made collages, using this technology to mitigate the relentless demands of the new digital theocracy.
Collage images are posted daily on Instagram and Twitter from Monday to Saturday.
Human collage: Monday, Wednesday, Friday
GAN collage: Tuesday, Friday, Saturday
Matteo Rattini, https://www.instagram.com/thisculpturedoesntexist/
Valentina Di Liscia, ‘The Art of AI: Computer-generated Sculptures Are Eerily Real’ , August 3, 2021, hyperallergic.com
Filippo Lorenzin , ‘Appreciating the Performative Quality of Computer Generated Art’ February 8, 2022, hyperallergic.com
Sarah Rose Sharp ‘DALL-E,” the New AI Artist Who Can Draw Anything’ April 13, 2022, hyperallergic.com
Astral Codex Ten, https://astralcodexten.substack.com/p/a-guide-to-asking-robots-to-design/
Google CoLab Pro
Find me on social media to see the results of the Digital Twin project.
The first GAN image was published on June 7th, 2022.
A Generative Adversarial Network (GAN) is a new method of machine learning where two neural networks contest to produce FAKE data from an initial REAL data (see diagram below). Like mimicry in human evolutionary biology, the ‘generator’ neural network trains to fool the ‘discriminator’ neural network, enabling the framework to ‘learn’ independently to produce an unlimited new FAKE image dataset from the REAL image dataset.