Abstract
This paper presents a novel harmony generation method based on decision-theoretic planning. We are the first to model music generation using Markov decision processes (MDPs). We give a proof of concept for this approach by using MDP planning to generate four-part harmony, given the melody or soprano line. Our initial results show feasibility, and show the variance possible, depending on the choice of reward functions.
Original language | English |
---|---|
Pages (from-to) | 263-274 |
Number of pages | 12 |
Journal | International Journal of Approximate Reasoning |
Volume | 51 |
Issue number | 2 |
DOIs | |
State | Published - Jan 2010 |
Bibliographical note
Funding Information:This work is partially supported by NSF Grant ITR-0325063.
Funding
This work is partially supported by NSF Grant ITR-0325063.
Funders | Funder number |
---|---|
National Science Foundation (NSF) | ITR-0325063 |
Directorate for Computer and Information Science and Engineering | 0325063 |
Keywords
- Decision-theoretic planning
- Harmony generation
- Markov decision processes
ASJC Scopus subject areas
- Software
- Artificial Intelligence
- Theoretical Computer Science
- Applied Mathematics