EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Random rotational embedding Bayesian optimization for human-in-the-loop personalized music generation

Miguel Marcos iD, Lorenzo Mur-Labadia, Ruben Martinez-Cantin iD

PLOS One · 2025

DOI: 10.1371/journal.pone.0335853 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, 2025. Verfasst von Miguel Marcos und 2 weiteren Autor:innen. Beteiligte Einrichtung: Departamento de Informática e Ingeniería de Sistemas, Instituto de Investigación en Ingeniería de Aragón (I3A), Universidad de Zaragoza, Zaragoza, Sp…. Hauptthemen laut MEDLINE: Music, Deep Learning. Die Arbeit nennt 81 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Agencia Estatal de Investigación, Departamento de Educación, Cultura y Deporte, Gobierno de Aragón.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Generative deep learning models, such as those used for music generation, can produce a wide variety of results based on perturbations of random points in their latent space. User preferences can be incorporated in the generative process by replacing this random sampling with a personalized query. Bayesian optimization, a sample-efficient nonlinear optimization method, is the gold standard for human-in-the-loop optimization problems, such as finding this query. In this paper, we present random rotational embedding Bayesian optimization (ROMBO). This novel method can efficiently sample and optimize high-dimensional spaces with rotational symmetries, like the Gaussian latent spaces found in generative models. ROMBO works by embedding a low-dimensional Gaussian search space into a high-dimensional one through random rotations. Our method outperforms several baselines, including other high-dimensional Bayesian optimization variants. We evaluate our algorithm through a music generation task. Our evaluation includes both simulated experiments and real user feedback. Our results show that ROMBO can perform efficient personalization of a generative deep learning model. The main contributions of our paper are: we introduce a novel embedding strategy for Bayesian optimization in high-dimensional Gaussian sample spaces; achieve a consistently better performance throughout optimization with respect to baselines, with a final loss reduction of 16%-31% in simulation; and complement our simulated evaluations with a study with human volunteers (n = 16). Users working with our music generation pipeline find new favorite pieces 40% more often, 16% faster, and spend 18% less time on pieces they dislike than when randomly querying the model. These results, along with a final survey, demonstrate great performance and satisfaction, even among users with particular tastes.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

MusicDeep LearningHumansBayes TheoremAlgorithms

Autor:innen und Institutionen

  • Departamento de Informática e Ingeniería de Sistemas, Instituto de Investigación en Ingeniería de Aragón (I3A), Universidad de Zaragoza, Zaragoza, Spain.

ORCID-Profile: Miguel Marcos, Ruben Martinez-Cantin

Förderung

  • Agencia Estatal de Investigación
  • Departamento de Educación, Cultura y Deporte, Gobierno de Aragón

Zitieren

APA

Marcos, M., Mur-Labadia, L., & Martinez-Cantin, R. (2025). Random rotational embedding Bayesian optimization for human-in-the-loop personalized music generation. PLOS One. https://doi.org/10.1371/journal.pone.0335853

Vancouver

Marcos M, Mur-Labadia L, Martinez-Cantin R. Random rotational embedding Bayesian optimization for human-in-the-loop personalized music generation. PLOS One. 2025. doi:10.1371/journal.pone.0335853

BibTeX
@article{Marcos2025,
  author = {Marcos, Miguel and Mur-Labadia, Lorenzo and Martinez-Cantin, Ruben},
  title = {Random rotational embedding Bayesian optimization for human-in-the-loop personalized music generation},
  journal = {PLOS One},
  year = {2025},
  doi = {10.1371/journal.pone.0335853}
}
Kostenlos starten →