Abstract
This paper outlines an approach to real-time music generation using melody and harmony focused agents in a process inspired by jazz improvisation. A harmony agent employs a Long Short-Term Memory (LSTM) artificial neural network trained on the chord progressions of 2986 jazz ‘standard’ compositions using a network structure novel to chord sequence analysis. The melody agent uses a rule-based system of manipulating provided, pre-composed melodies to improvise new themes and variations. The agents take turns in leading the direction of the composition based on a rating system that rewards harmonic consistency and melodic flow. In developing the multi-agent system it was found that implementing embedded spaces in the LSTM encoding process resulted in significant improvements to chord sequence learning.
| Original language | English |
|---|---|
| Title of host publication | Computational Intelligence in Music, Sound, Art and Design |
| Subtitle of host publication | 6th International Conference, EvoMUSART 2017, Proceedings |
| Editors | Joao Correia, Vic Ciesielski, Antonios Liapis |
| Place of Publication | Cham Switzerland |
| Publisher | Springer |
| Pages | 114-127 |
| Number of pages | 14 |
| Volume | 10198 |
| ISBN (Electronic) | 9783319557502 |
| ISBN (Print) | 9783319557496 |
| DOIs | |
| Publication status | Published - 2017 |
| Event | International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design 2017 - Amsterdam, Netherlands Duration: 19 Apr 2017 → 21 Apr 2017 Conference number: 6th https://link.springer.com/book/10.1007/978-3-319-55750-2 (proceedings) |
Publication series
| Name | Lecture Notes in Computer Science |
|---|---|
| Publisher | Springer |
| Volume | 10198 |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Conference
| Conference | International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design 2017 |
|---|---|
| Abbreviated title | EvoMUSART 2017 |
| Country | Netherlands |
| City | Amsterdam |
| Period | 19/04/17 → 21/04/17 |
| Internet address |
|
Keywords
- Artificial neural networks
- Multi-agent systems
- Music composition
Cite this
}
Using autonomous agents to improvise music compositions in Real-Time. / Hutchings, Patrick; McCormack, Jon.
Computational Intelligence in Music, Sound, Art and Design : 6th International Conference, EvoMUSART 2017, Proceedings. ed. / Joao Correia; Vic Ciesielski; Antonios Liapis . Vol. 10198 Cham Switzerland : Springer, 2017. p. 114-127 (Lecture Notes in Computer Science; Vol. 10198).Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review
TY - GEN
T1 - Using autonomous agents to improvise music compositions in Real-Time
AU - Hutchings, Patrick
AU - McCormack, Jon
PY - 2017
Y1 - 2017
N2 - This paper outlines an approach to real-time music generation using melody and harmony focused agents in a process inspired by jazz improvisation. A harmony agent employs a Long Short-Term Memory (LSTM) artificial neural network trained on the chord progressions of 2986 jazz ‘standard’ compositions using a network structure novel to chord sequence analysis. The melody agent uses a rule-based system of manipulating provided, pre-composed melodies to improvise new themes and variations. The agents take turns in leading the direction of the composition based on a rating system that rewards harmonic consistency and melodic flow. In developing the multi-agent system it was found that implementing embedded spaces in the LSTM encoding process resulted in significant improvements to chord sequence learning.
AB - This paper outlines an approach to real-time music generation using melody and harmony focused agents in a process inspired by jazz improvisation. A harmony agent employs a Long Short-Term Memory (LSTM) artificial neural network trained on the chord progressions of 2986 jazz ‘standard’ compositions using a network structure novel to chord sequence analysis. The melody agent uses a rule-based system of manipulating provided, pre-composed melodies to improvise new themes and variations. The agents take turns in leading the direction of the composition based on a rating system that rewards harmonic consistency and melodic flow. In developing the multi-agent system it was found that implementing embedded spaces in the LSTM encoding process resulted in significant improvements to chord sequence learning.
KW - Artificial neural networks
KW - Multi-agent systems
KW - Music composition
UR - http://www.scopus.com/inward/record.url?scp=85017535190&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-55750-2_8
DO - 10.1007/978-3-319-55750-2_8
M3 - Conference Paper
SN - 9783319557496
VL - 10198
T3 - Lecture Notes in Computer Science
SP - 114
EP - 127
BT - Computational Intelligence in Music, Sound, Art and Design
A2 - Correia, Joao
A2 - Ciesielski, Vic
A2 - Liapis , Antonios
PB - Springer
CY - Cham Switzerland
ER -