Finnish researchers unveil DeepBeat, a rhyme-generating hip-hop computer system.
A robot, probably writing raps. Image by Arthur Caranta via Flickr
These days, with all the rapping holograms, rapping grannies, and rapping tomatoes, you’d think there would be a wealth of fun rapping robots out there. But considering the fact that we already have robot chefs and robot journalists, there’s a shockingly low number of rhyming androids. Now, researchers at the University of Aalto in Finland are taking us one step closer to a world of mechanical MCs with “DeepBeat,” an algorithm-based, computerized system for making hip-hop happen. Rapping is a lot of hard work, you see, and like with many other difficult jobs, some scientists envision a future where we can kick back, and let computers do the rapping for us.
“The highly structured nature of rap makes it particularly amenable to computer analysis,” the MIT Technology Review writes. “And that raises an interesting question: if computers can analyze rap lyrics, can they also generate them?”
Now we know the answer is “yes, kind of.”
Image by Mike Licht via Flickr
In their paper, titled “DopeLearning: A Computational Approach to Rap Lyrics Generation,” the Finnish researchers report using a set of more than 10,000 songs from 100 different musicians to program DeepBeat, focusing on artists with high rhyming density, like Rakim and Wu Tang’s Inspectah Deck. They also taught it to recognize sequential information by subject matter—when prompted with a lyric from a song it knew, DeepBeat was able to pull the song’s real next line from a list of possible choices 82 percent of the time. Using this knowledge of rhyme and context, the algorithm was then put to the task of composing it’s own songs, using snippets of other tracks from existing artists. Here’s a little ditty it came up with on the subject of love, using lyrics from Big Daddy Kane, Lil’ Wayne, Eminem, and Snoop Dogg, among others:
For a chance at romance I would love to enhance
But everything I love has turned to a tedious task
One day we gonna have to leave our love in the past
I love my fans but no one ever puts a grasp
I love you momma, I love my momma, I love you momma
And I would love to have a thing like you on my team you take care
I love it when it's sunny Sonny girl you could be my Cher
I'm in a love affair I can't share it ain't fair
Haha I'm just playin' ladies you know I love you.
I know my love is true and I know you love me too
Girl I'm down for whatever cause my love is true
This one goes to my man Old Dirty one love we be swigging brew
My brother I love you be encouraged man and just know
When you done let me know cause my love make you be like WHOA
If I can't do it for the love then do it I won't
All I know is I love you too much to walk away, though
Not too bad for a computer. Well, maybe it’s the subject matter. Robots were never really any good at this crazy thing we humans call love. If these lyrics are typical, DeepBeat probably shouldn’t quit its day job yet. But the exercise is just one more way machine learning is moving into what we thought was the sole domain of human creativity. And though Kendrick Lamar probably won’t be replaced with an algorithm in the near future, DeepBeat shows enough potential that we can expect to see the weirdness, awkward hilarity, and cultural repercussions of more robot rap very soon.