Fluidic Abstraction
Machine-learning study generating visuals from the conceptual worlds of radical designers — Zaha Hadid, Archigram, Superstudio and more.
Fluidic Abstraction explores the possibilities of generating visuals based on conceptual works and ideas of selected designers — Zaha Hadid, Peter Cook, MAD, Archigram, Superstudio and others — from different eras and fields, but with similar or sometimes radical beliefs.
A machine-learning model is trained to examine and assimilate the distinct methodologies, conceptual frameworks and aesthetic details characteristic of these influential designers, with the aim of producing original visual outputs informed by those learnings.
Outputs
Each output sits somewhere between its sources: neither a copy of any single designer's language nor a random remix, but an interpolation across the latent space the model has learned.
