Finding Temporal Structure in Music: Blues Improvisation with LSTM
Recurrent Networks
D. Eck, and J. Schmidhuber. Neural Networks for Signal Processing XII, Proceedings of the 2002
IEEE Workshop, page 747--756. New York, IEEE, (2002)
Abstract
Few types of signal streams are as ubiquitous as music. Here we consider
the problem of extracting essential ingredients of music signals,
such as well-defined global temporal structure in the form of nested
periodicities (or meter). Can we construct an adaptive signal
processing device that learns by example how to generate new instances
of a given musical style? Because recurrent neural networks can in
principle learn the temporal structure of a signal, they are good
candidates for such a task. Unfortunately, music composed by standard
recurrent neural networks (RNNs) often lacks global coherence. The
reason for this failure seems to be that RNNs cannot keep track of
temporally distant events that indicate global music structure. Long
Short-Term Memory (LSTM) has succeeded in similar domains where other
RNNs have failed, such as timing & counting and learning of context
sensitive languages. In the current study we show that LSTM is also
a good mechanism for learning to compose music. We present experimental
results showing that LSTM successfully learns a form of blues music
and is able to compose novel (and we believe pleasing) melodies in
that style. Remarkably, once the network has found the relevant structure
it does not drift from it: LSTM is able to play the blues with good
timing and proper structure as long as one is willing to listen.
%0 Conference Paper
%1 Eck2002d
%A Eck, D.
%A Schmidhuber, J.
%B Neural Networks for Signal Processing XII, Proceedings of the 2002
IEEE Workshop
%C New York
%D 2002
%E Bourlard, H.
%I IEEE
%K imported
%P 747--756
%T Finding Temporal Structure in Music: Blues Improvisation with LSTM
Recurrent Networks
%X Few types of signal streams are as ubiquitous as music. Here we consider
the problem of extracting essential ingredients of music signals,
such as well-defined global temporal structure in the form of nested
periodicities (or meter). Can we construct an adaptive signal
processing device that learns by example how to generate new instances
of a given musical style? Because recurrent neural networks can in
principle learn the temporal structure of a signal, they are good
candidates for such a task. Unfortunately, music composed by standard
recurrent neural networks (RNNs) often lacks global coherence. The
reason for this failure seems to be that RNNs cannot keep track of
temporally distant events that indicate global music structure. Long
Short-Term Memory (LSTM) has succeeded in similar domains where other
RNNs have failed, such as timing & counting and learning of context
sensitive languages. In the current study we show that LSTM is also
a good mechanism for learning to compose music. We present experimental
results showing that LSTM successfully learns a form of blues music
and is able to compose novel (and we believe pleasing) melodies in
that style. Remarkably, once the network has found the relevant structure
it does not drift from it: LSTM is able to play the blues with good
timing and proper structure as long as one is willing to listen.
@inproceedings{Eck2002d,
abstract = {Few types of signal streams are as ubiquitous as music. Here we consider
the problem of extracting essential ingredients of music signals,
such as well-defined global temporal structure in the form of nested
periodicities (or {\em meter}). Can we construct an adaptive signal
processing device that learns by example how to generate new instances
of a given musical style? Because recurrent neural networks can in
principle learn the temporal structure of a signal, they are good
candidates for such a task. Unfortunately, music composed by standard
recurrent neural networks (RNNs) often lacks global coherence. The
reason for this failure seems to be that RNNs cannot keep track of
temporally distant events that indicate global music structure. Long
Short-Term Memory (LSTM) has succeeded in similar domains where other
RNNs have failed, such as timing \& counting and learning of context
sensitive languages. In the current study we show that LSTM is also
a good mechanism for learning to compose music. We present experimental
results showing that LSTM successfully learns a form of blues music
and is able to compose novel (and we believe pleasing) melodies in
that style. Remarkably, once the network has found the relevant structure
it does not drift from it: LSTM is able to play the blues with good
timing and proper structure as long as one is willing to listen.},
added-at = {2010-02-27T01:05:18.000+0100},
address = {New York},
author = {Eck, D. and Schmidhuber, J.},
biburl = {https://www.bibsonomy.org/bibtex/233c4a4cfd873c76a905046df7ab1106c/tb2332},
booktitle = {Neural Networks for Signal Processing XII, Proceedings of the 2002
IEEE Workshop},
editor = {Bourlard, H.},
file = {2002_ieee.pdf:papers/2002_ieee.pdf:PDF;2002_ieee.ps.gz:papers/2002_ieee.ps.gz:PostScript},
interhash = {36baf273a5e22466c2273d28e5c17e33},
intrahash = {33c4a4cfd873c76a905046df7ab1106c},
keywords = {imported},
owner = {thierry},
pages = {747--756},
publisher = {IEEE},
source = {OwnPublication},
timestamp = {2010-02-27T01:05:21.000+0100},
title = {Finding Temporal Structure in Music: Blues Improvisation with {LSTM}
Recurrent Networks},
year = 2002
}