English

BYOL for Audio: Self-Supervised Learning for General-Purpose Audio Representation

Audio and Speech Processing 2021-04-22 v2 Machine Learning Sound

Abstract

Inspired by the recent progress in self-supervised learning for computer vision that generates supervision using data augmentations, we explore a new general-purpose audio representation learning approach. We propose learning general-purpose audio representation from a single audio segment without expecting relationships between different time segments of audio samples. To implement this principle, we introduce Bootstrap Your Own Latent (BYOL) for Audio (BYOL-A, pronounced "viola"), an audio self-supervised learning method based on BYOL for learning general-purpose audio representation. Unlike most previous audio self-supervised learning methods that rely on agreement of vicinity audio segments or disagreement of remote ones, BYOL-A creates contrasts in an augmented audio segment pair derived from a single audio segment. With a combination of normalization and augmentation techniques, BYOL-A achieves state-of-the-art results in various downstream tasks. Extensive ablation studies also clarified the contribution of each component and their combinations.

Keywords

Cite

@article{arxiv.2103.06695,
  title  = {BYOL for Audio: Self-Supervised Learning for General-Purpose Audio Representation},
  author = {Daisuke Niizumi and Daiki Takeuchi and Yasunori Ohishi and Noboru Harada and Kunio Kashino},
  journal= {arXiv preprint arXiv:2103.06695},
  year   = {2021}
}

Comments

IJCNN 2021, 8 pages, 4 figures

R2 v1 2026-06-23T23:59:54.722Z