Automated composition of Galician Xota - tuning RNN-based composers for specific musical styles using Deep Q-Learning. Mira, R., Coutinho, E., Parada-Cabaleiro, E., & Schuller, B. PeerJ Computer Science, 9:e1356, 5, 2023. ![Automated composition of Galician Xota - tuning RNN-based composers for specific musical styles using Deep Q-Learning [link]](https://bibbase.org/img/filetypes/link.svg "link")
Website doi abstract bibtex Music composition is a complex field that is difficult to automate because the computational definition of what is good or aesthetically pleasing is vague and subjective. Many neural network-based methods have been applied in the past, but they lack consistency and in most cases, their outputs fail to impress. The most common issues include excessive repetition and a lack of style and structure, which are hallmarks of artificial compositions. In this project, we build on a model created by Magenta—the RL Tuner—extending it to emulate a specific musical genre—the Galician Xota. To do this, we design a new rule-set containing rules that the composition should follow to adhere to this style. We then implement them using reward functions, which are used to train the Deep Q Network that will be used to generate the pieces. After extensive experimentation, we achieve an implementation of our rule-set that effectively enforces each rule on the generated compositions, and outline a solid research methodology for future researchers looking to use this architecture. Finally, we propose some promising future work regarding further applications for this model and improvements to the experimental procedure.
@article{
title = {Automated composition of Galician Xota - tuning RNN-based composers for specific musical styles using Deep Q-Learning},
type = {article},
year = {2023},
keywords = {automated music composition,deep q-learning,galician xota,magenta,rl-tuner},
pages = {e1356},
volume = {9},
websites = {https://peerj.com/articles/cs-1356},
month = {5},
day = {15},
id = {dc6d199d-44fa-3ab0-b581-bebb3057fe16},
created = {2023-05-15T08:14:20.617Z},
file_attached = {true},
profile_id = {ffa9027c-806a-3827-93a1-02c42eb146a1},
last_modified = {2023-08-23T15:24:46.421Z},
read = {false},
starred = {false},
authored = {true},
confirmed = {true},
hidden = {false},
private_publication = {false},
abstract = {Music composition is a complex field that is difficult to automate because the computational definition of what is good or aesthetically pleasing is vague and subjective. Many neural network-based methods have been applied in the past, but they lack consistency and in most cases, their outputs fail to impress. The most common issues include excessive repetition and a lack of style and structure, which are hallmarks of artificial compositions. In this project, we build on a model created by Magenta—the RL Tuner—extending it to emulate a specific musical genre—the Galician Xota. To do this, we design a new rule-set containing rules that the composition should follow to adhere to this style. We then implement them using reward functions, which are used to train the Deep Q Network that will be used to generate the pieces. After extensive experimentation, we achieve an implementation of our rule-set that effectively enforces each rule on the generated compositions, and outline a solid research methodology for future researchers looking to use this architecture. Finally, we propose some promising future work regarding further applications for this model and improvements to the experimental procedure.},
bibtype = {article},
author = {Mira, R and Coutinho, Eduardo and Parada-Cabaleiro, E and Schuller, Björn},
doi = {10.7717/peerj-cs.1356},
journal = {PeerJ Computer Science}
}
Downloads: 0
{"_id":"m5CY2rkWP3W2RE6Rf","bibbaseid":"mira-coutinho-paradacabaleiro-schuller-automatedcompositionofgalicianxotatuningrnnbasedcomposersforspecificmusicalstylesusingdeepqlearning-2023","author_short":["Mira, R.","Coutinho, E.","Parada-Cabaleiro, E.","Schuller, B."],"bibdata":{"title":"Automated composition of Galician Xota - tuning RNN-based composers for specific musical styles using Deep Q-Learning","type":"article","year":"2023","keywords":"automated music composition,deep q-learning,galician xota,magenta,rl-tuner","pages":"e1356","volume":"9","websites":"https://peerj.com/articles/cs-1356","month":"5","day":"15","id":"dc6d199d-44fa-3ab0-b581-bebb3057fe16","created":"2023-05-15T08:14:20.617Z","file_attached":"true","profile_id":"ffa9027c-806a-3827-93a1-02c42eb146a1","last_modified":"2023-08-23T15:24:46.421Z","read":false,"starred":false,"authored":"true","confirmed":"true","hidden":false,"private_publication":false,"abstract":"Music composition is a complex field that is difficult to automate because the computational definition of what is good or aesthetically pleasing is vague and subjective. Many neural network-based methods have been applied in the past, but they lack consistency and in most cases, their outputs fail to impress. The most common issues include excessive repetition and a lack of style and structure, which are hallmarks of artificial compositions. In this project, we build on a model created by Magenta—the RL Tuner—extending it to emulate a specific musical genre—the Galician Xota. To do this, we design a new rule-set containing rules that the composition should follow to adhere to this style. We then implement them using reward functions, which are used to train the Deep Q Network that will be used to generate the pieces. After extensive experimentation, we achieve an implementation of our rule-set that effectively enforces each rule on the generated compositions, and outline a solid research methodology for future researchers looking to use this architecture. Finally, we propose some promising future work regarding further applications for this model and improvements to the experimental procedure.","bibtype":"article","author":"Mira, R and Coutinho, Eduardo and Parada-Cabaleiro, E and Schuller, Björn","doi":"10.7717/peerj-cs.1356","journal":"PeerJ Computer Science","bibtex":"@article{\n title = {Automated composition of Galician Xota - tuning RNN-based composers for specific musical styles using Deep Q-Learning},\n type = {article},\n year = {2023},\n keywords = {automated music composition,deep q-learning,galician xota,magenta,rl-tuner},\n pages = {e1356},\n volume = {9},\n websites = {https://peerj.com/articles/cs-1356},\n month = {5},\n day = {15},\n id = {dc6d199d-44fa-3ab0-b581-bebb3057fe16},\n created = {2023-05-15T08:14:20.617Z},\n file_attached = {true},\n profile_id = {ffa9027c-806a-3827-93a1-02c42eb146a1},\n last_modified = {2023-08-23T15:24:46.421Z},\n read = {false},\n starred = {false},\n authored = {true},\n confirmed = {true},\n hidden = {false},\n private_publication = {false},\n abstract = {Music composition is a complex field that is difficult to automate because the computational definition of what is good or aesthetically pleasing is vague and subjective. Many neural network-based methods have been applied in the past, but they lack consistency and in most cases, their outputs fail to impress. The most common issues include excessive repetition and a lack of style and structure, which are hallmarks of artificial compositions. In this project, we build on a model created by Magenta—the RL Tuner—extending it to emulate a specific musical genre—the Galician Xota. To do this, we design a new rule-set containing rules that the composition should follow to adhere to this style. We then implement them using reward functions, which are used to train the Deep Q Network that will be used to generate the pieces. After extensive experimentation, we achieve an implementation of our rule-set that effectively enforces each rule on the generated compositions, and outline a solid research methodology for future researchers looking to use this architecture. Finally, we propose some promising future work regarding further applications for this model and improvements to the experimental procedure.},\n bibtype = {article},\n author = {Mira, R and Coutinho, Eduardo and Parada-Cabaleiro, E and Schuller, Björn},\n doi = {10.7717/peerj-cs.1356},\n journal = {PeerJ Computer Science}\n}","author_short":["Mira, R.","Coutinho, E.","Parada-Cabaleiro, E.","Schuller, B."],"urls":{"Website":"https://peerj.com/articles/cs-1356"},"biburl":"https://bibbase.org/service/mendeley/ffa9027c-806a-3827-93a1-02c42eb146a1","bibbaseid":"mira-coutinho-paradacabaleiro-schuller-automatedcompositionofgalicianxotatuningrnnbasedcomposersforspecificmusicalstylesusingdeepqlearning-2023","role":"author","keyword":["automated music composition","deep q-learning","galician xota","magenta","rl-tuner"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/ffa9027c-806a-3827-93a1-02c42eb146a1","dataSources":["2252seNhipfTmjEBQ"],"keywords":["automated music composition","deep q-learning","galician xota","magenta","rl-tuner"],"search_terms":["automated","composition","galician","xota","tuning","rnn","based","composers","specific","musical","styles","using","deep","learning","mira","coutinho","parada-cabaleiro","schuller"],"title":"Automated composition of Galician Xota - tuning RNN-based composers for specific musical styles using Deep Q-Learning","year":2023}