0

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

BYOL-A, an audio self-supervised learning method, generates general-purpose audio representations from single audio segments without relying on time-based audio relationships, achieving state-of-the-art results in various downstream tasks.

Year
2021
Venue
arXiv 2021
Authors
5
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2103.06695v2ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

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.

Authors

5