An international team of computer scientists and musicologists are determined to complete Beethoven’s 10th symphony from fragmentary notes left by the genius using AI for the 250th anniversary celebrations of his birth in 2020.

German-based international team is led by the director of the Karajan Institute in Salzburg Matthias Röder, who is known for founding the Karajan Music Tech Conference and being a proponent for hackathons in the music industry.

Röder believes that “our human creativity will be enhanced by machines and humanity will reach a new level of creativity both on an individual and, more importantly, on a collective basis”. While first results were probably far too simplistic, now he admits that AI really surprises him:

“An AI system learns an unbelievable amount of notes in an extremely short time. And the first results are a bit like with people, you say ‘hmm, maybe it’s not so great’. But it keeps going and, at some point, the system really surprises you.”

Machine learning algorithm is being trained not only on the remaining notes but also on other examples of Ludwig van Beethoven’s work and other composers’ compositions that are believed to be influential to his work.

The project is sponsored by Deutsche Telekom, that bets on using research proceeding for further R&D in fields such as voice recognition.

It is not the first project of the kind. Perhaps most famous among them was dedicated to completing Schubert’s 10th symphony. While critics were not much impressed with the results at the time, with AI growth rate reportedly beating Moore’s law, this time AI can probably measure to human-level performance in one more field.