Data-efficient self-supervised algorithms for fine-grained birdsong analysis

Research in bioacoustics, neuroscience, and linguistics often uses birdsong as a proxy to acquire knowledge across diverse areas. This requires audio models to annotate and parse the birdsong. Developing such models requires precise, syllable-level annotated training data. Therefore, automated metho...

Full description

Bibliographic Details

Published in:	Ecological Informatics
Main Authors:	Houtan Ghaffari, Lukas Rauch, Paul Devos
Format:	Article
Language:	English
Published:	Elsevier 2026-06-01
Subjects:	Birdsong analysis Deep learning Self-supervised learning
Online Access:	http://www.sciencedirect.com/science/article/pii/S1574954126002682