Time-Shifted Token Scheduling for Symbolic Music Generation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wang, Ting-Kang, Tan, Chih-Pin, Yang, Yi-Hsuan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912612660281344
author Wang, Ting-Kang
Tan, Chih-Pin
Yang, Yi-Hsuan
author_facet Wang, Ting-Kang
Tan, Chih-Pin
Yang, Yi-Hsuan
contents Symbolic music generation faces a fundamental trade-off between efficiency and quality. Fine-grained tokenizations achieve strong coherence but incur long sequences and high complexity, while compact tokenizations improve efficiency at the expense of intra-token dependencies. To address this, we adapt a delay-based scheduling mechanism (DP) that expands compound-like tokens across decoding steps, enabling autoregressive modeling of intra-token dependencies while preserving efficiency. Notably, DP is a lightweight strategy that introduces no additional parameters and can be seamlessly integrated into existing representations. Experiments on symbolic orchestral MIDI datasets show that our method improves all metrics over standard compound tokenizations and narrows the gap to fine-grained tokenizations.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23749
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Time-Shifted Token Scheduling for Symbolic Music Generation
Wang, Ting-Kang
Tan, Chih-Pin
Yang, Yi-Hsuan
Machine Learning
Symbolic music generation faces a fundamental trade-off between efficiency and quality. Fine-grained tokenizations achieve strong coherence but incur long sequences and high complexity, while compact tokenizations improve efficiency at the expense of intra-token dependencies. To address this, we adapt a delay-based scheduling mechanism (DP) that expands compound-like tokens across decoding steps, enabling autoregressive modeling of intra-token dependencies while preserving efficiency. Notably, DP is a lightweight strategy that introduces no additional parameters and can be seamlessly integrated into existing representations. Experiments on symbolic orchestral MIDI datasets show that our method improves all metrics over standard compound tokenizations and narrows the gap to fine-grained tokenizations.
title Time-Shifted Token Scheduling for Symbolic Music Generation
topic Machine Learning
url https://arxiv.org/abs/2509.23749