mlsanityData to 14 Oct 2026

Beyond task performance: Decoding bioacoustic embeddings with speech features

arXiv 2606.14662 ↗First posted 12 Jun 2026cs.LG cs.SDcited by 0

Not on the map. Why?

Abstract (arXiv)

Pretrained audio embeddings are standard in bioacoustics, yet little is known about which acoustic features these models encode, nor which are useful for a given task. This hinders transparency and limits extension to rare species or data-scarce domains. Here we reveal which speech-like features are encoded in bioacoustic representations. Using the 88~eGeMAPS features across six taxonomic groups, we apply linear and nonlinear regression probes to quantify which acoustic properties each model captures. Results confirm a ``no free lunch'' pattern: no single model captures the full feature space. A concatenated embedding achieves the highest performance, suggesting complementary acoustic space coverage across models. Loudness features are best encoded (R2 = 0.76) while F0 is hardest to recover (R2 = 0.33). By cross-referencing recoverability with per-species feature salience (NMI), we derive data-driven model selection guidance for bioacoustics.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 20 of its 35 references that are in the map. First posted on arXiv 12 Jun 2026.

Builds on

20 of 35 references.

References · 20

  1. 2017Deep Image Priorcited by 1,098
  2. 2021Probing Classifiers: Promises, Shortcomings, and Advancescited by 876
  3. 2021SUPERB: Speech processing Universal PERformance Benchmarkcited by 666
  4. 2022BEATs: Audio Pre-Training with Acoustic Tokenizerscited by 388
  5. 2021Layer-wise Analysis of a Self-supervised Speech Representation Modelcited by 298
  6. 2022Comparative layer-wise analysis of self-supervised speech modelscited by 158
  7. 2021Computational bioacoustics with deep learning: a review and roadmapcited by 78
  8. 2022HEAR: Holistic Evaluation of Audio Representationscited by 78
  9. 2023Global birdsong embeddings enable superior transfer learning for bioacoustic classificationcited by 51
  10. 2022BEANS: The Benchmark of Animal Soundscited by 38
  11. 2024NatureLM-audio: an Audio-Language Foundation Model for Bioacousticscited by 33
  12. 2025Can Masked Autoencoders Also Listen to Birds?cited by 26
8 more
  1. 2022Opening the Black Box of wav2vec Feature Encodercited by 18
  2. 2025Foundation Models for Bioacoustics -- a Comparative Reviewcited by 11
  3. 2025Clustering and novel class recognition: evaluating bioacoustic deep learning feature extractorscited by 7
  4. 2025Towards a Unified Representation Evaluation Framework Beyond Downstream Taskscited by 7
  5. 2025Crossing the Species Divide: Transfer Learning from Speech to Animal Soundscited by 4
  6. 2024Probing the Information Encoded in Neural-based Acoustic Models of Automatic Speech Recognition Systemscited by 1
  7. 2026Embryonic Exposure to VPA Influences Chick Vocalisations: A Computational Studycited by 1
  8. 2026Multi-layer attentive probing improves transfer of audio representations for bioacousticscited by 1

Built on by

Not cited yet.