Machine Learning for Music Makers

Machine learning already plays a role in many technologies that we encounter on a daily basis. Relatively speaking, however, the field is still in its infancy, with plenty of untapped potential – including for the creative industries.

This workshop introduces some of the ways machine learning can be applied to music. We’ll broadly survey the landscape of techniques and research relevant to both music making – production, synthesis, and composition – as well as recommendation engines and playlist generation. To get some hands-on experience, we’ll analyze a large custom set of audio files with the librosa audio analysis library, and arrange them spatially using the t-SNE dimensionality reduction technique.

No previous knowledge of machine learning is required for this workshop. Software and example audio sets will be provided, but please feel free to bring your own music to use in the practical section.