Music Creation with Machine Learning

People

Supervisor

Research areas

Description

This project will involve using machine learning (ML) techniques to create some type of musical data. This could involve learning about and creating some kind of sequence learning models (e.g., the LSTM-RNNs or a Transformer), and then training this model on a dataset of your choice (e.g., digital audio data or "symbolic music" data in MIDI format). You could also use different styles of ML, such as reinforcement learning, or regression.

There are a wide variety of research questions to be explored about how different representations of music and different kinds of musical datasets work with different ML techniques. Part of this project will involve choosing an area that interests you and developing your research planning skills!

As a first step, you'll look at the Creative Prediction project tutorials to learn the basics of music generation with ML.

Goals

  • Learn about content generation with Machine Learning
  • Apply a machine learning models to musical datasets
  • Create new music with a machine learning system
  • Critically examine the results.

Requirements

  • Enthusiasm for musical topics and creative arts.
  • Demonstrated coursework experience with machine learning OR coursework in computer music/ creative computing.

Keywords

machine learning, AI, music, generation, creative prediction.

Updated:  1 June 2019/Responsible Officer:  Head of School/Page Contact:  CECS Webmaster