Can you tell the difference between Bach and Computer?
It’s not the first time that scientists attempt to substitute human composers with an Artificial Intelligence.
We already examined the attempt of a team of researchers from Kingston University and Queen Mary University of London, teaching an AI system how to compose, analyzing and learning from more than 23,000 Irish folk songs.
Now the stakes are higher. No more just Irish folk songs, now the reference model comes from Bach.
DeepBach is a deep learning-powered program created by Gaetan Hadjeres and François Pachet of the Sony Computer Science Laboratories in Paris.
These guys developed a model of polyphonic music generation, which learns to compose chorales in the style of Bach.
Here an example:
That’s how they explain that:
DeepBach, a graphical model aimed at modeling polyphonic music and specifically hymn-like pieces. We claim that, after being trained on the chorale harmonizations by Johann Sebastian Bach, our model is capable of generating highly convincing chorales in the style of Bach. DeepBach's strength comes from the use of pseudo-Gibbs sampling coupled with an adapted representation of musical data. This is in contrast with many automatic music composition approaches which tend to compose music sequentially. Our model is also steerable in the sense that a user can constrain the generation by imposing positional constraints such as notes, rhythms or cadences in the generated score. We also provide a plugin on top of the MuseScore music editor making the interaction with DeepBach easy to use.
Another example: God Save the Queen I
So, they applied an algorithm to Bach’s work. Why him?
First of all because he was incredibly prolific, that means many samples to train the AI. Second, he’s the master of the well-structured Baroque music, made of various and intricate patterns.
The point now is, does the student surpass the teacher?
Tests showed that, when given a DeepBach-generated harmony, around half the voters judged that it was composed by Bach.
“We consider this to be a good score knowing the complexity of Bach’s compositions,” say Hadjeres and Pachet.
Even when confronted with music composed by Bach himself, participants only judged that correctly 75 percent of the time.
Where will AI bring music technology?