David J. Robinson

David Robinson, Adelaide Robinson, Lily Akrapongpisak

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, Marius Miron, Masato Hagiwara et al.

Large language models (LLMs) prompted with text and audio have achieved state-of-the-art performance across various auditory tasks, including speech, music, and general audio, showing emergent abilities on unseen tasks. However, their potential has yet to be fully demonstrated in bioacoustics tasks, such as detecting animal vocalizations in large recordings, classifying rare and endangered species, and labeling context and behavior -- tasks that are crucial for conservation, biodiversity monitoring, and animal behavior studies. In this work, we present NatureLM-audio, the first audio-language foundation model specifically designed for bioacoustics. Our training dataset consists of carefully curated text-audio pairs spanning bioacoustics, speech, and music, designed to address the field's limited availability of annotated data. We demonstrate successful transfer of learned representations from music and speech to bioacoustics, and our model shows promising generalization to unseen taxa and tasks. We evaluate NatureLM-audio on a novel benchmark (BEANS-Zero) and it sets a new state of the art on several bioacoustics tasks, including zero-shot classification of unseen species. To advance bioacoustics research, we release our model weights, benchmark data, and open-source the code for training and benchmark data generation and model training.