LO-10: Computational Music

Author: Christine McLeavey Payne (An assortment of people created the actual product)
Date: April 25, 2019
Project: MuseNet

MuseNet is a neural network that can develop music of its own. It can create short compositions in the style of different artists/composers with instruments of choice. This is a very creative idea because it allows people to compose music in the style of someone like Beethoven, but with guitar instead. I admire the creativity in the project, but also this combination of so many different styles and types of music. Much of what is produced by MuseNet is compositions that would never be possible or done without the computational work done by the team. The algorithms were created by a neural network, essentially a type of code that learns based off of inputs. In doing so, the computer learned patterns from music rather than individual analysis having to be done by people for every type of music/song. The artistic sensibilities are shown by the creative aspect of MuseNet. Clearly, the team cared about music and wanted to create a unique twist so they used their computer science knowledge to do this.

Link to Blog: https://openai.com/blog/musenet/

Link to Example: https://soundcloud.com/openai_audio/pop-nocturne