MachineLearning_button_hover

Machine Learning Related Work

Disco Cube [2020]

In collaboration with visual artist Rommelo Yu, this work was developed using a deep learning solution for generating sculpture geometries, which borrows from the field of AI-robot task completion and Hamiltonian Cycle graph theory. We investigate a chain of directives that are passed from (i) Rommelo to me, (ii) from me to my computer, (iii) and finally from my computer back to Rommelo. The final chain of communication is a set of instructions with which Rommelo will construct a 4m^2 sculpture of aluminum and glass (a Disco Cube) that will hang in the night club, Berghain - in celebration of the Ostgut Ton label’s 15th anniversary. The sculpture is cubic, and of cubic dimension 15x15.

Understanding Morph Targets (UMT)  [2019]

Developed to address a compositional desire to synthesize vocal content with the rhythm and cadence of ‘break-beats’ for use in electronic music. Using Machine Learning and MIR techniques, this software recreates a ‘target’ audiofile using a temporal reconfiguration of a ‘source’ audiofile. The tonal and rhythmic qualities of the target are played out using only the audio material from the ‘source’. This technique can further morph on levels of complexity that surpass basic music theory, where cultural-specific styles and tonalities can be merged to create new musical expressions.

Works produced using UMT have been presented on Noods Radio, Cashmere Radio, and live at Stray Signals #7 in Berlin. 

The German publishing house, S. Fischer Verlag, has commissioned a new work to celebrate the publication of Olivia Wenzel’s novel, 1000 Serpantinen Angst. Created using UMT, the work will morph the musical culture of the protagonist's demographic in the story, with that of Wenzel, and will score a short film released alongside the novel.