Plenty of rappers have claimed that they’re the best alive. But now they’ve got some competition—from a computer program.
A team of Finnish computer scientists has developed a machine-learning algorithm, dubbed “DeepBeat,” which has “learned” how to rap by analyzing the rhymes present in approximately 10,000 songs by more than 100 artists.
When it came to creating the program, the scientists chose to focus exclusively on a type of linguistic device known as an assonance rhyme—which, unlike a “perfect rhyme,” only requires that the words in question repeat vowel sounds—because of their prevalence in rap music.
Once DeepBeat figured out how to pick out assonance rhymes, it could be programmed to choose subsequent lines to match. (Eighty-two percent of the time, it successfully picked the line that actually follows in the song it had mined.) But it could also be instructed to search through its database for a verse that fits a given line’s rhyme structure and is about the same topic.
DeepBeat put together this romantic tune using rhyming lines from a variety of different hits:
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
The verses it produces are occasionally incoherent, but the Massachusetts Institute of Technology's Technology Review notes that DeepBeat’s rhymes are significantly more complex than those produced by human rappers, who are outperformed “in terms of length and frequency of the rhymes,” the researchers write, “21 percent of the time.”
The programmers have also used DeepBeat to rank the rhyming skills of the rappers whose work their algorithm has studied. Inspector Deck, Rakim, and Redrama came out on top; Eminem, on the other hand, didn’t fare so well. But according to Technology Review, “That’s probably because Eminem often achieves his rhymes by ‘bending’ words, a trick that this technique does not allow for.”
The team has yet to teach DeepBeat to consistently produce flows containing coherent narratives, which means Jay Z, Marshall Mathers, and Lil Wayne's legacies are safe. For now.
[h/t Technology Review]