acm-header
Sign In

Communications of the ACM

ACM TechNews

Facebook AI Turns One Style of Music Into Another


View as: Print Mobile App Share:
A robot orchestra.

Facebook's AI Research unit has developed a neural network that can translate music from one style, genre, and set of instruments into another.

Credit: Nicole Gray

Facebook AI Research (FAIR) has developed a neural network that can translate music from one style, genre, and set of instruments into another.

The artificial intelligence (AI) system takes one input, such as a symphony orchestra playing Bach, and translates it into another version of the same song, such as on a piano in the style of Beethoven.

The FAIR team says this marks the first time an unsupervised learning method has been able to recreate high-fidelity music with a neural network.

The new approach involves a complex method of auto-encoding that enables the network to process audio from inputs on which it has never been trained. The unsupervised learning method uses high-level semantics interpretation, instead of attempting to match pitch or memorize notes.

From The Next Web
View Full Article

 

Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account