Frosty Morning [Dataset] 2019 Resolution: 1600x1200 1:14 loop
A dataset video from the installation, Frosty Morning (2019)
Frosty Morning is a multi-channel video installation generated by Neural Networks, which are rhythmic structures made by human beings that simulate our brain and nervous systems. After the artist’s research on Machine Learning and Neural Networks, an Artificial Neural Network is trained to “read” the front page of The New York Times from the last 6 years. By imitating the data set, it starts to render the newspaper, generate blurred image, and approach the form of reality, but never realizing the legible content. As a perpetual newspaper generator or a machine trying to predict “the future,” the work seems to reflect the reasonable and overwhelming world in the present.
Aesthetically, the moving image generated by AI somehow is reminiscent of abstract landscape paintings. From a technical perspective, the landscape corroded by AI is what we call the “latent space,” the high-dimensional space that is invisible between the form of history and memory. In Frosty Morning, the passage of time brought by the moving image has transferred this concept to the “latent time” which is outside of human time and bigger than the nature time.
http://shi-zheng.net/works-frosty_morning.html