Project name: Machine learning + Art


Another experimental project 'Machine learning Art'. Creative freedom on obsessive colours of the universe. Addition to my 'data+art' collective. Inspecting beauty in data and depicting the complexity of computational thinking. Method of amalgamation of Nebula image library from NASA's Hubble Space Telescope trained on StyleGAN model and experimenting 3d heightmaps on the same.


Data, artificial intelligence, machine learning (ML), and deep learning (DL): These make up the zeitgeist of our current times. StyleGAN is a novel generative adversarial network introduced by Nvidia researchers in December 2018, and open sourced in February 2019. StyleGAN depends on Nvidia's CUDA software, GPUs and on TensorFlow.

I started my career with no cultural history of art. Always pushing on technology hard for creating my designs/art. But also have the least respect for generative works made without the intent involving an artistic intent created by human beings. But any work of art with intent remains relevant irrespective of the medium as long it is appreciated by at least one man among the audience.

The project also aim to evoke a sense of other-worldliness with its generative particles with added turbulence. This is a road to depict alien graphics apart from the usual 'cyberpunk' with the help of unusual physics simulation and higher dimensional geometries. 


​The idea of ‘storytelling’ in projections is one way forward in terms of content beyond the simple projection of images and shapes. There is the opportunity to engage people in stories beyond looking at a building and delve into history and stories and use the format with a little more narrative around histories. With this, it’s possible to take data from almost anything and use that to create an animation which is live.

Barcode of colors — Visualization of color data used in the StyleGAN video of the space images

(c) 2020