Embodied / Divergent 2024
Embodied / Divergent is an interactive digitial artwork that connects human touch with the latent architecture of an AI system. As visitors use the touch interface, their interaction subverts and shapes the AI’s internal processes – disrupting its usual behaviour and revealing unexpected visial outcomes. By making the invisible mechanics of machine intelligence tangiable, the installation invites audiences to interact with a system that is usually hidden.
The installation uses network bending techniques – a toolkit developed by Terence Broad – which modify and manipulate the computational graph of a generative model. Originally used for GANs, part of my master’s dissertation project at UAL’s Creative Computing Institute was to implement these techniques into a diffusion model and allow for real-time control over them. The interface I designed for this project acts as a layer of abstraction for these techniques, which often require a deep understanding of the model’s architecture, and lets the user control the model internally using an XY pad. The interface also includes an interactive U-Net map, allowing the user to select different layers in the downsampling block to apply them to.
Collection of Artworks
The installation was further adapted for London Tech Week 2025 to enable human movement as a means of interaction with its internal processes. Positional data from a Kinect was mapped using regression models, translating XY coordinates into multiple network bending parameters for real-time control. Additionally, the position of a second person was mapped to determine which layer of the U-Net the parameters were applied to, enabling collaborative interaction and dynamic modulation of the model’s internal structure.
Audiovisual Work
Below is an audiovisual piece made using the same techniques. Music written and performed by me.
https://garincurtis.com