Hit enter to search or ESC to close
Members of the Australian band Uncanny Valley trained an algorithm using sounds from koalas.
Uncanny Valley Music, Sound, and Technology
By Rodrigo Pérez Ortega Apr. 24, 2020 , 3:30 PM
A contest inspired by the popular annual Eurovision music competition has drawn entries composed with the help of unusual songwriters—artificial intelligence (AI) programs.
For the first time, 13 teams from Europe and Australia, made up of data scientists, programmers, and musicians, submitted tunes this month to the AI Song Contest, sponsored by Dutch public broadcaster VPRO. According to the contest’s website, the organizers seek to raise awareness of the important role of AI in our lives, as well as the possibilities and limitations of the technology.
The contestants were invited to write AI programs or adapt existing ones to craft lyrics, melodies, and harmonies for songs lasting no more than 3 minutes each. Besides entering their tracks, the teams also submitted their algorithms, models, and code, along with explanations of their creative process and the systems they used.
The Australian team Uncanny Valley, created “Beautiful The World” by training a neural network with past Eurovision songs and audio from koalas, kookaburras, and Tasmanian devils to honor those animals lost to the country’s wildfires in 2019. The team’s AI algorithm was in charge of coming up with the melody and lyrics, and they added human vocals to the song.
A three-person panel—two AI scientists and a composer—and the public will rate the songs (vote here) by 10 May, with results to be announced on 12 May. The winning song will be broadcast on national Dutch radio.
The annual Eurovision competition that inspired the AI contest was canceled for this year because of the COVID-19 pandemic, but European broadcasters are planning alternative shows in May to replace what would have been the awards ceremony.
About the author
Your email address will not be published. Required fields are marked *
Save my name, email, and website in this browser for the next time I comment.