Select papers and resources for the NYU GenAudio research reading groups that I have led.
NYU GenAudio regularly hosts weekly research paper reading groups encouraging people to read research on topics ranging from generative audio, machine learning, signal processing, and music technology. I have presented research from various disciplines in an approachable light that encourages anybody (regardless of background) to engage with the material. The notes, slides, and research papers that I have used are available on my GitHub. This page details the different research papers that I have presented at GenAudio.
HarmonyCloak: Making Music Unlearnable for Generative AI
HarmonyCloak is a paper published by the MOSIS lab from the University of Tennessee, Knoxville’s department of Electrical Engineering and Computer Science. It describes an adversarial noise injection attack that prevents Generative Audio models (like Suno or Udio) from learning off a piece of music (MIDI or Audio file)
The Concatenator: A Bayesian Approach to Concatenative Musaicing
The Concatenator was written by Chris Tralie (ex-Ursinus College, DSP & Computer Graphics Researcher) and Ben Cantil (CTO Datamind Audio). It describes a real-time algorithm for concatenative synthesis. This algorithm appears in the audio plugin of the same name by Datamind Audio.