“Generative Adversarial Network (millimeter wave)” by Matthew Biederman


  • ©, Matthew Biederman, Generative Adversarial Network (millimeter wave)

Title:


    Generative Adversarial Network (millimeter wave)

Artist(s) and People Involved:


Symposium:


Venue(s):



Artist Statement:


    Focus Québec Exhibition. Forum des Images, May 16 – 21

    A Generative Adversarial Network
    This work takes its name from the software framework used to create its imagery as a self-referential point in order to reveal not only how the images are created, but also how the idea of the public has shifted in a so-called free society. The parameters of what constitutes an adversary are constantly shifting depending on the context and perspective, but within machinic perception, they remain fixed within the biases of its programmers and the dataset, or society and their polity. For ‘A Generative Adversarial Network’ a machine-learning algorithm was trained with a dataset gleaned from millimeter-wave security scans. Rather than using it for security screening as intended, the algorithm creates new images of imaginary people. This methodology of AI creates new images from random noise by continually updating and re-evaluating the imagery, at times breaking down, becoming confused and starting over – not unlike our global security apparatus.


Other Information:



Website:



Additional Images: