Refik Anadol: Unsupervised — Machine Hallucinations at MoMA
Primary Source: https://www.moma.org/calendar/exhibitions/5535
Author: Refik Anadol · Year: 2022–2023 · Publisher: Museum of Modern Art (MoMA)
Summary
Documentation of Refik Anadol’s monumental installation ‘Unsupervised’ at the Museum of Modern Art (MoMA). Using custom machine learning models trained on 200 years of MoMA’s digitized art collection, the installation visualizes continuous fluid navigations through multi-dimensional latent space in real time.
Key Points
- Latent Space Cartography: Training generative neural networks on museum archives to treat historical art metadata as a continuous mathematical vector space.
- Real-Time Environmental Modulation: Incorporating live sensory data (sound, movement, weather) to steer the generative model’s real-time trajectory.
- Data Aesthetics & Machine Hallucination: Transforming static archival datasets into fluid kinetic pigment and architectural-scale projections.