Using tropical reef, bird and unrelated sounds for superior transfer learning in marine bioacoustics
Philosophical Transactions of the Royal Society B: Biological Sciences · 12.06.2025 · 380 (1928)
DOI: 10.1098/rstb.2024.0280 ↗Kurz gefasst
Fachartikel in der Zeitschrift Philosophical Transactions of the Royal Society B: Biological Sciences, online erschienen am 12.06.2025. Verfasst von Ben Williams und 14 weiteren Autor:innen. Beteiligte Einrichtung: University College London, London, UK. Hauptthemen laut MEDLINE: Birds, Coral Reefs, Machine Learning, Acoustics. Die Arbeit wurde bisher 39-mal zitiert (Crossref) und nennt 57 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fisheries Society of the British Isles, American Elasmobranch Society.
Abstract (Originalsprache)
Machine learning has the potential to revolutionize passive acoustic monitoring (PAM) for ecological assessments. However, high annotation and computing costs limit the field’s adoption. Generalizable pretrained networks can overcome these costs, but high-quality pretraining requires vast annotated libraries, limiting their current development to data-rich bird taxa. Here, we identify the optimum pretraining strategy for data-deficient domains, using tropical reefs as a representative case study. We assembled ReefSet, an annotated library of 57 000 reef sounds taken across 16 datasets, though still modest in scale compared to annotated bird libraries. We performed multiple pretraining experiments and found that pretraining on a library of bird audio 50 times the size of ReefSet provides notably superior generalizability on held-out reef datasets, with a mean area under the receiver operating characteristic curve (AUC-ROC) of 0.881 (±0.11), compared to pretraining on ReefSet itself or unrelated audio, with a mean AUC-ROC of 0.724 (±0.05) and 0.834 (±0.05), respectively. However, our key findings show that cross-domain mixing, where bird, reef and unrelated audio are combined during pretraining, provides superior transfer learning performance, with an AUC-ROC of 0.933 (±0.02). SurfPerch, our optimum pretrained network, provides a strong foundation for automated analysis of tropical reef and related PAM data with minimal annotation and computing costs. This article is part of the theme issue ‘Acoustic monitoring for tropical ecology and conservation’.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- University College London, London, UK.
- Institute of Zoology, Zoological Society of London, London, UK.
- DeepMind, Google Inc, Mountain View, CA, USA.
- DeepMind, Google Inc, Canada.
- DeepMind, Google Inc, New York, USA.
- Conservation Metrics, Santa Cruz, CA, USA.
ORCID-Profile: Ben Williams, Abram B Fleishman, Aaron N Rice, David Curnick, Kate E Jones
Förderung
- Fisheries Society of the British Isles
- American Elasmobranch Society
Zitieren
APA
Williams, B., van Merriënboer, B., Dumoulin, V., Hamer, J., Fleishman, A. B., McKown, M., Munger, J., Rice, A. N., Lillis, A., White, C., Hobbs, C., Razak, T., Curnick, D., Jones, K. E., & Denton, T. (2025). Using tropical reef, bird and unrelated sounds for superior transfer learning in marine bioacoustics. Philosophical Transactions of the Royal Society B: Biological Sciences, 380(1928). https://doi.org/10.1098/rstb.2024.0280
Vancouver
Williams B, van Merriënboer B, Dumoulin V, Hamer J, Fleishman AB, McKown M, et al. Using tropical reef, bird and unrelated sounds for superior transfer learning in marine bioacoustics. Philosophical Transactions of the Royal Society B: Biological Sciences. 2025;380(1928). doi:10.1098/rstb.2024.0280
BibTeX
@article{Williams2025,
author = {Williams, Ben and van Merriënboer, Bart and Dumoulin, Vincent and Hamer, Jenny and Fleishman, Abram B and McKown, Matthew and Munger, Jill and Rice, Aaron N and Lillis, Ashlee and White, Clemency and Hobbs, Catherine and Razak, Tries and Curnick, David and Jones, Kate E and Denton, Tom},
title = {Using tropical reef, bird and unrelated sounds for superior transfer learning in marine bioacoustics},
journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
year = {2025},
volume = {380},
number = {1928},
doi = {10.1098/rstb.2024.0280}
}
EUVIMED