Marimba-Playing Robot Composes Its Own Music Using AI and Deep Learning

A Georgia Tech robot named Shimon is using deep learning and a database of over 5,000 songs - including works Beethoven, the Beatles, and Miles Davis, among others - to compose its own music. The researchers also fed the bot more than two million motifs and riffs of music. Aside from giving the machine the first four measures to use as a starting point, no humans are involved in either the composition or the performance of the music. Ph.D. student Mason Bretan is the man behind the machine. He's worked with Shimon for seven years, enabling it to 'listen' to music played by humans and improvise over pre-composed chord progressions. Now Shimon is a solo composer for the first time, generating the melody and harmonic structure on its own.

Transcript

00:00:00 (marimba music) - We've done a lot of stuff where Shimon is interacting with people, and improvising over predetermined chord progressions, but here, Shimon is generating everything by itself. There's no human influence, other than we give it a C at the very beginning, and then we say, Shimon, generate the rest of the composition from here. And then Shimon takes it away, and completes the piece.

00:00:44 We've trained the neural network with a database that consists of a wide variety of music, encompassing everything from jazz, classical, to pop music. What it's generating is almost like a fusion of these different styles. What you're hearing is the result of what it's learned about music in general. And so, because it's learned all these concepts, that sort of crossover between genres,

00:01:10 the output is a mixture of all of these different things.