The Merger of Machine Learning, Music and Programming
summary
Claims concrete results: unconditional and conditional audio samples generated with MusicGen and MusicLM style models, vocal cloning that convincingly imitates commercial artists, and stem separation that produces usable a cappella tracks. Demo code uses Python libraries — Pedalboard, librosa, Transformers and diffusers — served as microservices via MLServer, with a working frontend that generates, visualizes waveforms/spectrograms, applies effects, and returns playable files; the demos mostly work though prompt tuning remains trial and error.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 2:38Music programming & ML landscape
- 7:53How music becomes digital audio
- 11:52Why music matters in science
- 14:00Music programming languages
- 17:38Machine learning fundamentals and its role
- 20:09ML applications in music
- 27:27Live demo: AI-generated music
- 30:20Microservices for music models
- 43:55Serving models with ML Server
- 48:35Adding audio effects
- 50:08Closing remarks and key takeaways
related talks
41mMusic and Machine Learning
30mSound Modelling - parametric methods and deep learning...
36mMusic and Art with ML
58mMusic Generation via Masked Acoustic Token Modeling
43mMaking machines that make music
54mIntegral scherzo: how machines compose and sing
01h 04mKeynote 3. Interactive generative models for music
46mMusic Generation with Magenta: Using Machine Learning in