Linux, Objects
Leave a comment

Animal Classifier – Arbitrary classifications for machine learning

Created by Seoul based artistic duo Shinseungback Kimyonghun, Animal Classifier is an AI trained to divide animals into arbitrary classifications to foreground the imperfections and edge cases in classification systems.

Taxonomy was used as derived from the essay ‘The Analytical Language of John Wilkins’ by Jorge Luis Borges where he discusses arbitrarities of John Wilkins and writes – “it is clear that there is no classification of the Universe not being arbitrary and full of conjectures. The reason for this is very simple: we do not know what thing the universe is.”.

“Most of the current AI systems are basically classifiers, and they learn and work based on the classifications provided by humans, thus inevitably imperfect.” – Shinseungback Kimyonghun

The installation uses TensorFlow, Inception-v3, online images from Flickr, 14 Raspberry PIs, 14 engraved metal plates, 14 glass domes and 14 monitors. The full list of images used for classification by the AI is available here.

Project Page | Shinseungback Kimyonghun