A Contrastive Self-Supervised Learning scheme for beat tracking amenable to few-shot learning - Département Image, Données, Signal
Conference Papers Year : 2024

A Contrastive Self-Supervised Learning scheme for beat tracking amenable to few-shot learning

Abstract

In this paper, we propose a novel Self-Supervised-Learning scheme to train rhythm analysis systems and instantiate it for few-shot beat tracking. Taking inspiration from the Contrastive Predictive Coding paradigm, we propose to train a Log-Mel-Spectrogram Transformer encoder to contrast observations at times separated by hypothesized beat intervals from those that are not. We do this without the knowledge of ground-truth tempo or beat positions, as we rely on the local maxima of a Predominant Local Pulse function, considered as a proxy for Tatum positions, to define candidate anchors, candidate positives (located at a distance of a power of two from the anchor) and negatives (remaining time positions). We show that a model pre-trained using this approach on the unlabeled FMA, MTT and MTG-Jamendo datasets can successfully be fine-tuned in the few-shot regime, i.e. with just a few annotated examples to get a competitive beat-tracking performance.
Fichier principal
Vignette du fichier
ISMIR2024_template.pdf (747.46 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04768296 , version 1 (05-11-2024)

Identifiers

  • HAL Id : hal-04768296 , version 1

Cite

Antonin Gagnere, Geoffroy Peeters, Slim Essid. A Contrastive Self-Supervised Learning scheme for beat tracking amenable to few-shot learning. ISMIR 2024 : 25th International Society for Music Information Retrieval (ISMIR) Conference Onsite and Virtual, Nov 2024, San Francisco, Californ, United States. ⟨hal-04768296⟩
0 View
0 Download

Share

More