Home BlogDataset Teaching Machines to Understand Classical Music

Teaching Machines to Understand Classical Music

by Joshua New
Classical music

Researchers at the University of Washington have published a dataset named MusicNet consisting of 330 highly annotated classical music recordings to help machine learning researchers train algorithms to understand aspects of classical music. MusicNet’s recordings have over one million total annotations, indicating the time of every note, the instrument used to play each note, and other useful information about song structure. By training on MusicNet, machine learning algorithms could better understand aspects of classical music to support applications ranging from music transcription to song recommendation.

Get the data.

Image: FotoshopTofs

You may also like

Show Buttons
Hide Buttons