Writing rap lyrics requires both creativity to construct a meaningful,
interesting story and lyrical skills to produce complex rhyme patterns, which
form the cornerstone of good flow. We present a rap lyrics generation method
that captures both of these aspects. First, we develop a prediction model to
identify the next line of existing lyrics from a set of candidate next lines.
This model is based on two machine-learning techniques: the RankSVM algorithm
and a deep neural network model with a novel structure. Results show that the
prediction model can identify the true next line among 299 randomly selected
lines with an accuracy of 17%, i.e., over 50 times more likely than by random.
Second, we employ the prediction model to combine lines from existing songs,
producing lyrics with rhyme and a meaning. An evaluation of the produced lyrics
shows that in terms of quantitative rhyme density, the method outperforms the
best human rappers by 21%. The rap lyrics generator has been deployed as an
online tool called DeepBeat, and the performance of the tool has been assessed
by analyzing its usage logs. This analysis shows that machine-learned rankings
correlate with user preferences.
Description
DopeLearning: A Computational Approach to Rap Lyrics Generation
%0 Generic
%1 malmi2015dopelearning
%A Malmi, Eric
%A Takala, Pyry
%A Toivonen, Hannu
%A Raiko, Tapani
%A Gionis, Aristides
%D 2015
%K analysis deep generating kallimachos learning lyric svm text
%R 10.1145/2939672.2939679
%T DopeLearning: A Computational Approach to Rap Lyrics Generation
%U http://arxiv.org/abs/1505.04771
%X Writing rap lyrics requires both creativity to construct a meaningful,
interesting story and lyrical skills to produce complex rhyme patterns, which
form the cornerstone of good flow. We present a rap lyrics generation method
that captures both of these aspects. First, we develop a prediction model to
identify the next line of existing lyrics from a set of candidate next lines.
This model is based on two machine-learning techniques: the RankSVM algorithm
and a deep neural network model with a novel structure. Results show that the
prediction model can identify the true next line among 299 randomly selected
lines with an accuracy of 17%, i.e., over 50 times more likely than by random.
Second, we employ the prediction model to combine lines from existing songs,
producing lyrics with rhyme and a meaning. An evaluation of the produced lyrics
shows that in terms of quantitative rhyme density, the method outperforms the
best human rappers by 21%. The rap lyrics generator has been deployed as an
online tool called DeepBeat, and the performance of the tool has been assessed
by analyzing its usage logs. This analysis shows that machine-learned rankings
correlate with user preferences.
@misc{malmi2015dopelearning,
abstract = {Writing rap lyrics requires both creativity to construct a meaningful,
interesting story and lyrical skills to produce complex rhyme patterns, which
form the cornerstone of good flow. We present a rap lyrics generation method
that captures both of these aspects. First, we develop a prediction model to
identify the next line of existing lyrics from a set of candidate next lines.
This model is based on two machine-learning techniques: the RankSVM algorithm
and a deep neural network model with a novel structure. Results show that the
prediction model can identify the true next line among 299 randomly selected
lines with an accuracy of 17%, i.e., over 50 times more likely than by random.
Second, we employ the prediction model to combine lines from existing songs,
producing lyrics with rhyme and a meaning. An evaluation of the produced lyrics
shows that in terms of quantitative rhyme density, the method outperforms the
best human rappers by 21%. The rap lyrics generator has been deployed as an
online tool called DeepBeat, and the performance of the tool has been assessed
by analyzing its usage logs. This analysis shows that machine-learned rankings
correlate with user preferences.},
added-at = {2016-12-21T17:34:46.000+0100},
author = {Malmi, Eric and Takala, Pyry and Toivonen, Hannu and Raiko, Tapani and Gionis, Aristides},
biburl = {https://www.bibsonomy.org/bibtex/2db41b88e98a84156bd82c0c997f7647d/hotho},
description = {DopeLearning: A Computational Approach to Rap Lyrics Generation},
doi = {10.1145/2939672.2939679},
interhash = {0e23432bb99895793fa853f1dd9a80e2},
intrahash = {db41b88e98a84156bd82c0c997f7647d},
keywords = {analysis deep generating kallimachos learning lyric svm text},
note = {cite arxiv:1505.04771Comment: This is a pre-print of an article appearing at KDD'16},
timestamp = {2016-12-21T17:34:46.000+0100},
title = {DopeLearning: A Computational Approach to Rap Lyrics Generation},
url = {http://arxiv.org/abs/1505.04771},
year = 2015
}