Frosty Morning Daily
2021
Multi-channel Video Installation
Dimensions variable, portrait orientation
8 min, loop

Frosty Morning (2018) took the artist's research on machine learning and neural networks as a point of departure. In his practice, an artificial neural network was trained through sequence of frames of video to predict and generate images of the future. The artist collected front pages of The New York Times from the last six years, and transformed them into a dataset in chronological order to train the neural network. During the process, he reflected on data as well as the linear narrative of images. After extensive iterative learning, the neural network evolved into a "newspaper generator" that could infinitely produce contents, beyond comprehension though. The linear arrangement of the dataset and the root design of the systematic framework determined that the nature of the neural network's actions was trying to predict the future.

As a continuance of the Frosty Morning series, this time the artist installs the moving images of The New York Times generated by the neural network in computer-simulated scene of daily life. The scenes mostly originate from pictures in his collection. In this particular context, the newspaper, despite being the visual core of the work, seems more like a daily scene that we encounter inadvertently.