Exploring Music Generation with Magenta

Abstract

Magenta is a research project developed by Google that explores the role of machine learning in creating art and music. Primarily this involves developing new deep learning and reinforcement learning algorithms (Magenta 2020a). In addition, Magenta provides tools for managing and manipulating music/image data. This project will focus on using various Magenta models, specifically RNNs and autoencoders, which take symbolic music data as input to generate new samples. The goal is to give musicians another outlet for creativity, allowing them to create variations on existing MIDI data or generate new data.

Faculty Mentor: Dr. Dana Cobzas