Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Transferable Models for Bioacoustics with Human Language Supervision

About

Passive acoustic monitoring offers a scalable, non-invasive method for tracking global biodiversity and anthropogenic impacts on species. Although deep learning has become a vital tool for processing this data, current models are inflexible, typically cover only a handful of species, and are limited by data scarcity. In this work, we propose BioLingual, a new model for bioacoustics based on contrastive language-audio pretraining. We first aggregate bioacoustic archives into a language-audio dataset, called AnimalSpeak, with over a million audio-caption pairs holding information on species, vocalization context, and animal behavior. After training on this dataset to connect language and audio representations, our model can identify over a thousand species' calls across taxa, complete bioacoustic tasks zero-shot, and retrieve animal vocalization recordings from natural text queries. When fine-tuned, BioLingual sets a new state-of-the-art on nine tasks in the Benchmark of Animal Sounds. Given its broad taxa coverage and ability to be flexibly queried in human language, we believe this model opens new paradigms in ecological monitoring and research, including free-text search on the world's acoustic monitoring archives. We open-source our models, dataset, and code.

David Robinson, Adelaide Robinson, Lily Akrapongpisak• 2023

Related benchmarks

TaskDatasetResultRank
Text-to-Audio RetrievaliNatSounds 2024 (val)
mAP@1000 (Amphi)39.06
5
Text-to-Audio RetrievalSoundscape Datasets (HSN, NES, SNE, UHH, PER, SSW) 2024 (OOD)
HSN Score27.85
5
ClassificationSoundscape OOD datasets (HSN, NES, SNE, UHH, PER, SSW) inat-sounds 2024 (test)
HSN Score19.69
3
Soundscape ClassificationHSN
Accuracy19.69
2
Soundscape ClassificationNES
Accuracy32.18
2
Soundscape ClassificationPER
Accuracy7.68
2
Soundscape ClassificationUHH
Accuracy16.65
2
Text-to-Audio RetrievalPER
mAP@10009.73
2
Text-to-Audio RetrievalUHH
mAP@100032
2
Audio-to-image retrievalCBI
mAP51.44
2
Showing 10 of 21 rows

Other info

Follow for update