A Computer-Generated Christmas Carol

To celebrate the holiday season, ASO is bringing you a science-filled 12 Days of ASO Christmas! This will be a series of ‘mini-blogs’, by Rosemary Wills, centered on the science of some of our most cherished traditions. This is the first in the series. 

Step aside, Bing Crosby. Computer scientists from the University of Toronto have created a program capable of composing a song to accompany any photo, using a technique they call “neural karaoke.” In the spirit of the holidays, they decided to feed their program a Christmas-y image of gifts under a tree.

image2

Image credit: Toby Bradbury via Flickr.

The result (available for listening here) is… frankly, a little weird. But you can’t deny it’s festive. Elfin voice? Check. Mention of a Christmas tree? Check. Themes of joy and hope? Possibly- it depends on how much poetic license you’re willing to give the lyrics. “The best Christmas present in the world is a blessing” seems to be on the right track, at least.

It may not bring peace on Earth (“I can hear music coming from the hall”? “I’ve always been there for the rest of our lives”?), but it does demonstrate an unusual use of artificial intelligence beyond informing you of the nearest pizza place to your location or, you know, taking over the world.

Here’s wishing you a holiday filled with “lots and lots and lots of flowers” (we’ll assume this means poinsettias).

About the Author

Rosemary WillisRosemary Wills is an undergraduate at UGA majoring in Plant Biology and Science Education. When she’s not writing, coding, or spending time with family, she enjoys growing plants in her windowsill and crocheting science-related things. More from Rosemary Wills.