The robot with four arms and eight sticks writes and plays its own compositions on a marimba, using a database of well-known pop, classical and jazz artists.
Researchers at Georgia Institute of Technology in the US fed the robot nearly 5,000 complete songs - from Beethoven to the Beatles to Lady Gaga to Miles Davis.
They worked with the robot named 'Shimon' for seven years, enabling it to listen to music played by humans and improvise over pre-composed chord progressions.
"Shimon's compositions represent how music sounds and looks when a robot uses deep neural networks to learn everything it knows about music from millions of human-made segments," said Mason Bretan, a PhD student at Georgia Tech.
More From This Section
"Once Shimon learns the four measures we provide, it creates its own sequence of concepts and composes its own piece," said Bretan.
"This is a leap in Shimon's musical quality because it's using deep learning to create a more structured and coherent composition," said Gil Weinberg, a professor at Georgia Tech.