Communication Dans Un Congrès Année : 2025

Taking Advantage of Emergent Properties of a Transformer for Music Representation Learning

Résumé

Self-supervised learning (SSL) has proven effective for music information retrieval (MIR), yet contrastive and equivariant methods excel on different tasks: contrastive learning on tagging, equivariant learning on structured predictions. We propose a multi-class-token multitask model (MT2) using multiple class tokens to optimize both contrastive and equivariant pretext tasks simultaneously, taking advantage of the emergent properties of a transformer. Averaging class tokens improves performance across diverse MIR tasks, while sequence tokens capture local musical information using emergent properties in a transformer. Our model achieves competitive results with far fewer parameters than existing masked modeling approaches.

Fichier principal
Vignette du fichier
kong2025cmmr.pdf (198 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05433459 , version 1 (28-12-2025)

Licence

Identifiants

  • HAL Id : hal-05433459 , version 1

Citer

Yuexuan Kong, Vincent Lostanlen, Romain Hennequin, Mathieu Lagrange, Gabriel Meseguer-Brocal. Taking Advantage of Emergent Properties of a Transformer for Music Representation Learning. Digital Music Research Network (DMRN), Dec 2025, Londres, United Kingdom. ⟨hal-05433459⟩
493 Consultations
115 Téléchargements

Partager

  • More