4. Thesis: Machine Memory










Immersive Environment for Explainable AI
Unreal Engine, Latent Terrain, nn tilde, MaxMSP, Blender, Ableton
Collaboration with the Sound Art Museum, Beijing

On going


Changing the inner structure of Stable Audio Open 1.0 (audio diffusion VAE model) to create new musical affordances under the theme of collective memory and nostalgia.

The project uses machine learning as a creative tool for musicians, creating effects for sounds in a similar manner to music production.

Sounds are collected by the Sound Art Museum, Beijing, and further manipulated into something new but still recognisable. 

The user interface is a ruins landscape made in Unreal Engine, with 3D scans of the local area.