Comparing Probabilistic Models for Melodic Sequences

Athina Spiliopoulou*, Amos Storkey

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

Modelling the real world complexity of music is a challenge for machine learning. We address the task of modeling melodic sequences from the same music genre. We perform a comparative analysis of two probabilistic models; a Dirichlet Variable Length Markov Model (Dirichlet-VMM) and a Time Convolutional Restricted Boltzmann Machine (TC-RBM). We show that the TC-RBM learns descriptive music features, such as underlying chords and typical melody transitions and dynamics. We assess the models for future prediction and compare their performance to a VMM, which is the current state of the art in melody generation. We show that both models perform significantly better than the VMM, with the Dirichlet-VMM marginally outperforming the TC-RBM. Finally, we evaluate the short order statistics of the models, using the Kullback-Leibler divergence between test sequences and model samples, and show that our proposed methods match the statistics of the music genre significantly better than the VMM.

Original languageEnglish
Title of host publicationMACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES, PT III
EditorsD Gunopulos, T Hofmann, D Malerba, M Vazirgiannis
Place of PublicationBERLIN
PublisherSpringer
Pages289-304
Number of pages16
ISBN (Electronic)978-3-642-23808-6
ISBN (Print)978-3-642-23807-9
DOIs
Publication statusPublished - 2011
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) - Athens, Greece
Duration: 5 Sept 20119 Sept 2011

Publication series

NameLecture Notes in Artificial Intelligence
PublisherSPRINGER-VERLAG BERLIN
Volume6913
ISSN (Print)0302-9743

Conference

ConferenceEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD)
Country/TerritoryGreece
Period5/09/119/09/11

Keywords / Materials (for Non-textual outputs)

  • melody modeling
  • music feature extraction
  • time convolutional restricted Boltzmann machine
  • variable length Markov model
  • Dirichlet prior
  • LEARNING ALGORITHM

Fingerprint

Dive into the research topics of 'Comparing Probabilistic Models for Melodic Sequences'. Together they form a unique fingerprint.

Cite this