ISBN 88-7395-155-4 © 2006 ICMPC 1848
Alma Mater Studiorum University of Bologna, August 22-26 2006
Improving algorithmic music composition with machine
learning
Michael Chan
School of Computer Science &
Engineering
University of New South Wales
Sydney, Australia
mchan@cse.unsw.edu.au
John Potter
School of Computer Science &
Engineering
University of New South Wales
Sydney, Australia
Emery Schubert
School of Music &
Music Education
University of New South Wales
Sydney, Australia
ABSTRACT
Algorithmic generation of musical sounding music is an
interesting but challenging task, because machines do not
inherently possess any form of creativity, which is neces-
sary to create music. The Automated Composer of Style
Sensitive Music II (ACSSM II) system generates music by
searching for a sequence of music segments that best sat-
isfy various constraints, including length and pitch range,
harmonic backbone, and consistency with a probabilistic
model of a composer’s style. As with all optimization prob-
lems, our problem requires the construction of a search
space; we utilize a clustering space produced by grouping
together music segments having similar musical features.
The output sequence is simply a path passing through these
clusters. In order to produce such a sequence, we utilize a
genetic algorithm. To evaluate the system, we have con-
ducted an experiment, involving five subjects who possess
at least three years of musical training, in which the over-
all musical quality of produced music was assessed. Our
results show that the automatically generated music
achieved a mean satisfaction score of 7.5/10, which is sig-
nificantly higher than that given to the music produced by
the earlier ACSSM system. Hence, the results suggest that
ACSSM II is a better system than its predecessor and is
capable of generating reasonably musical sounding music.
Keywords
Computer music, algorithmic music composition, artificial
intelligence, machine learning, clustering, genetic algo-
rithm, Markov chain, generative theory, musical intelli-
gence.
INTRODUCTION
One of the main goals of artificial intelligence (AI) is to
develop systems that imitate human intelligence and behav-
ior, and its success inspires researchers to focus on making
machines also imitate human creativity. One approach is to
develop a system that generates quality music, with little or
no supervision from the user. If such an automated system
could create musical sounding music, it might then be use-
ful in various ways; for example, human composers could
seek inspiration for their own compositions. More gener-
ally, from a computer science perspective, such a system
would represent a breakthrough in the application of AI
technology.
Our system, called the Automated Composer of Style-
Sensitive Music II (ACSSM II), aims to automate music
composition by exploiting an input corpus of music and
incorporating machine-learning techniques, including ge-
netic algorithms (Goldberg, 1989) and Markov chains
(Brémaud, 1999). It is based on an earlier system ACSSM
(Chan and Potter, 2005) which it extends with more exten-
sive AI techniques. Both of these systems have a common
fundamental concept, which is to generate new-sounding,
stylish music by “intelligently” rearranging segments of
music found in the input corpus. One interesting aspect of
In: M. Baroni, A. R. Addessi, R. Caterina, M. Costa (2006) Proceedings
of the 9th International Conference on Music Perception & Cognition
(ICMPC9), Bologna/Italy, August 22-26 2006.©2006 The Society for
Music Perception & Cognition (SMPC) and European Society for the
Cognitive Sciences of Music (ESCOM). Copyright of the content of an
individual paper is held by the primary (first-named) author of that pa-
per. All rights reserved. No paper from this proceedings may be repro-
duced or transmitted in any form or by any means, electronic or me-
chanical, including photocopying, recording, or by any information
retrieval systems, without permission in writing from the paper's primary
author. No other part of this proceedings may be reproduced or transmit-
ted in any form or by any means, electronic or mechanical, including
photocopying, recording, or by any information retrieval system, without
permission in writing from SMPC and ESCOM.