Knowledge-Driven Multi-Voice Music Generation
International Journal of Knowledge Management · 04.02.2026 · 22 (1) · 1-25
DOI: 10.4018/ijkm.400562 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Knowledge Management, erschienen am 04.02.2026. Verfasst von Chunquan Bian. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0).
Abstract (Originalsprache)
The rapid progress of artificial intelligence has advanced deep learning in music generation, particularly for polyphonic compositions where balancing harmony and coherence across multiple voices is crucial. Traditional approaches often face limitations such as rigid patterns, low flexibility, and difficulty optimizing global structures. This study introduces a multi-voice music generation framework that integrates an improved transformer with modulation and inversion-based data augmentation, enhancing harmony, coherence, controllability, and generalization. Experiments on the FMA dataset showed a 15.44% decrease in tone centrality index and a 20.69% reduction in polyphonic dissonance compared to the baseline transformer. User ratings indicated an 18.4% improvement in harmony and an 18.9% improvement in coherence. Under complex conditions, style consistency was 88.60%. These findings highlight the potential of AI-driven models to enhance creative knowledge processes, offering innovative pathways for music generation and knowledge-based digital creativity.
Zitieren
APA
Bian, C. (2026). Knowledge-Driven Multi-Voice Music Generation. International Journal of Knowledge Management, 22(1), 1-25. https://doi.org/10.4018/ijkm.400562
Vancouver
Bian C. Knowledge-Driven Multi-Voice Music Generation. International Journal of Knowledge Management. 2026;22(1):1-25. doi:10.4018/ijkm.400562
BibTeX
@article{Bian2026,
author = {Bian, Chunquan},
title = {Knowledge-Driven Multi-Voice Music Generation},
journal = {International Journal of Knowledge Management},
year = {2026},
volume = {22},
number = {1},
pages = {1--25},
doi = {10.4018/ijkm.400562}
}
EUVIMED