Machine Learning Art Projects by J. Rosenbaum

Machine learning art and projects in development. These projects comprise my experimentations and research into machine learning art and the research for my PhD exploring computer perceptions of gender. My artificial intelligence art centers around gender as its key theme and aims to increase representation of gender minorities in AI art, but also explores representation and bias in image classification and everyday uses of machine learning where bias may interfere with people outside the gender binary. Click on the image for further information about each of these projects.

An artwork of a transgender figure classically posed on a plinth and looking at their phone. This artwork has an augmented overlay produced using AI style transfer.
Beyond the Binary
A series of artworks exploring passing and non-passing transness. Made with 3D modeling, AI style transfer and Augmented Reality, these works are named after pronouns and use laser cutting, prints and AR to explore the depth beneath the surface.
An art print of bark that has been warped into an abstract image by a deep neural network
A series of works exploring environmental concerns through photography, 3D modeling, artificial intelligence and augmented reality.
An image generated using artificial intelligence art generator technique known as a GAN this work resembles a stone sculpture of a body that is crumbling.
Hidden Worlds
Using a Generative Adversarial Network and augmented reality along with a neural network generating captions for the works this augmented reality installation explores gender through the lens of computer vision.
Promotional image for Uncanny Valley, a hand holds a phone up to a machine learning GAN generated face. The phone is taking a selfie where the persons face has been overlaid with the face in the machine learning art.
Uncanny Valley
A look at representation and masking Using generative adversarial networks to generate images of faces and then using facial recognition and detection to apply those to people’s faces in Snapchat.