Synthetic data enables context-aware bioacoustic sound event detection
Not on the map. Why?
Abstract (arXiv)
We propose a methodology for training foundation models that enhances their in-context learning capabilities within the domain of bioacoustic signal processing. We use synthetically generated training data, introducing a domain-randomization-based pipeline that constructs diverse acoustic scenes with temporally strong labels. We generate over 8.8 thousand hours of strongly-labeled audio and train a query-by-example, transformer-based model to perform few-shot bioacoustic sound event detection. Our second contribution is a public benchmark of 13 diverse few-shot bioacoustics tasks. Our model outperforms previously published methods, and improves relative to other training-free methods by 64%. We demonstrate that this is due to increase in model size and data scale, as well as algorithmic improvements. We make our trained model available via an API, to provide ecologists and ethologists with a training-free tool for bioacoustic sound event detection.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 42 of its 75 references that are in the map. First posted on arXiv 1 Mar 2025.
Builds on
42 of 75 references.
References · 42
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2020Language Models are Few-Shot Learnerscited by 32,199
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
- 2017Prototypical Networks for Few-shot Learningcited by 3,001
- 2023FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioningcited by 2,551
- 2020GLU Variants Improve Transformercited by 1,724
- 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,401
- 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 616
- 2022BEATs: Audio Pre-Training with Acoustic Tokenizerscited by 378
- 2023Voicebox: Text-Guided Multilingual Universal Speech Generation at Scalecited by 373
- 2022Is synthetic data from generative models ready for image recognition?cited by 320
30 more
- 2022Images Speak in Images: A Generalist Painter for In-Context Visual Learningcited by 317
- 2018Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomizationcited by 292
- 2023WavCaps: A ChatGPT-Assisted Weakly-Labelled Audio Captioning Dataset for Audio-Language Multimodal Researchcited by 266
- 2022Visual Prompting via Image Inpaintingcited by 248
- 2023StableRep: Synthetic Images from Text-to-Image Models Make Strong Visual Representation Learnerscited by 177
- 2019Well-Read Students Learn Better: On the Importance of Pre-training Compact Modelscited by 118
- 2021Computational bioacoustics with deep learning: a review and roadmapcited by 78
- 2023SALM: Speech-augmented Language Model with In-context Learning for Speech Recognition and Translationcited by 75
- 2021Understanding Domain Randomization for Sim-to-real Transfercited by 69
- 2022AVES: Animal Vocalization Encoder based on Self-Supervisioncited by 53
- 2023Global birdsong embeddings enable superior transfer learning for bioacoustic classificationcited by 51
- 2023Towards In-context Scene Understandingcited by 45
- 2023Explore the Power of Synthetic Data on Few-shot Object Detectioncited by 38
- 2024NatureLM-audio: an Audio-Language Foundation Model for Bioacousticscited by 32
- 2021Improving Bird Classification with Unsupervised Sound Separationcited by 25
- 2023Transferable Models for Bioacoustics with Human Language Supervisioncited by 20
- 2023Learning to detect an animal sound from five examplescited by 14
- 2022Few-shot bioacoustic event detection at the DCASE 2022 challengecited by 13
- 2023Multimodal Foundation Models for Zero-shot Animal Species Recognition in Camera Trap Imagescited by 13
- 2023Unsupervised classification to improve the quality of a bird song recording datasetcited by 10
- 2023AnuraSet: A dataset for benchmarking Neotropical anuran calls identification in passive acoustic monitoringcited by 10
- 2023Can Self-Supervised Neural Representations Pre-Trained on Human Speech distinguish Animal Callers?cited by 9
- 2020Training Sound Event Detection On A Heterogeneous Datasetcited by 7
- 2023Regularized Contrastive Pre-training for Few-shot Bioacoustic Sound Detectioncited by 7
- 2024Mind the Domain Gap: a Systematic Analysis on Bioacoustic Sound Event Detectioncited by 7
- 2021A Mutual learning framework for Few-shot Sound Event Detectioncited by 3
- 2022Segment-level Metric Learning for Few-shot Bioacoustic Event Detectioncited by 3
- 2020Performance of a Deep Neural Network at Detecting North Atlantic Right Whale Upcallscited by 2
- 2024Synth4Kws: Synthesized Speech for User Defined Keyword Spotting in Low Resource Environmentscited by 2
- 2021Sound Event Detection Based on Curriculum Learning Considering Learning Difficulty of Eventscited by 1
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2025State Space Models for Bioacoustics: A Comparative Evaluation with Transformers
- 2025Advancing Marine Bioacoustics with Deep Generative Models: A Hybrid Augmentation Strategy for Southern Resident Killer Whale Detection
- 2024Self-supervised Learning for Acoustic Few-Shot Classification
- 2025SynSonic: Augmenting Sound Event Detection through Text-to-Audio Diffusion ControlNet and Effective Sample Filtering
- 2024Generative AI-based data augmentation for improved bioacoustic classification in noisy environments
- 2023DiffSED: Sound Event Detection with Denoising Diffusion
- 2025Knowledge-Augmented Vision Language Models for Underwater Bioacoustic Spectrogram Analysis
- 2025EvMic: Event-based Non-contact sound recovery from effective spatial-temporal modeling