A neural network completed Beethoven’s Tenth Symphony

Artificial intelligence was able to complete the Tenth Symphony by Ludwig van Beethoven, using for this purpose extant sketches and drafts of the composer.

Florian Colombo, a programmer and researcher from the École Polytechnique Fédérale de Lausanne, worked on the creation of a neural network called BeethovANN 10.1. It was she who helped finish the unfinished work, and conductor Guillaume Berni made some harmonic edits, after which the symphony was performed by the Nexus orchestra.

“It’s a pleasure to watch. It’s like being in a maternity ward and welcoming a new life,” Bernie shared.

Earlier, scientists used artificial intelligence to restore one of Rembrandt’s major works, “The Night Watch,” 300 years later. The painter finished the painting in 1642, and in 1715 the canvas was cropped on all four sides, so it would fit between the two doors of Amsterdam’s City Hall.

Specialists were able to add missing scenes and characters to the artwork. For this they used a copy of the painting and a model that can imitate the style of the famous painter as closely as possible.