Perch 2.0: The Bittern Lesson for Bioacoustics

• URL: http://arxiv.org/abs/2508.04665v1
• Date: Wed, 06 Aug 2025 17:34:43 GMT
• Title: Perch 2.0: The Bittern Lesson for Bioacoustics
• Abstract summary: Perch is a performant pre-trained model for bioacoustics.<n>It was trained in supervised fashion, providing both off-the-shelf classification scores for thousands of vocalizing species as well as strong embeddings for transfer learning.<n>In this new release, Perch 2.0, we expand from training exclusively on avian species to a large multi-taxa dataset.
• Score: 7.540404494602192
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Perch is a performant pre-trained model for bioacoustics. It was trained in supervised fashion, providing both off-the-shelf classification scores for thousands of vocalizing species as well as strong embeddings for transfer learning. In this new release, Perch 2.0, we expand from training exclusively on avian species to a large multi-taxa dataset. The model is trained with self-distillation using a prototype-learning classifier as well as a new source-prediction training criterion. Perch 2.0 obtains state-of-the-art performance on the BirdSet and BEANS benchmarks. It also outperforms specialized marine models on marine transfer learning tasks, despite having almost no marine training data. We present hypotheses as to why fine-grained species classification is a particularly robust pre-training task for bioacoustics.

Related papers

• Perch 2.0 transfers 'whale' to underwater tasks [6.216539628488371]
  Perch 2.0 is a supervised bioacoustics foundation model pretrained on 14,597 species.<n>Given that Perch 2.0 includes almost no marine mammal audio or classes in the training data, we evaluate Perch 2.0 performance on marine mammal and underwater audio tasks.
  arXiv  Detail & Related papers  (2025-12-02T20:49:41Z)
• Foundation Models for Bioacoustics -- a Comparative Review [0.9109149174920012]
  We review bioacoustic foundation models by thoroughly analysing design decisions such as model architecture, pretraining scheme, and training paradigm.<n>We evaluate selected foundation models on classification tasks from the BEANS and BirdSet benchmarks.<n>Our comprehensive experimental analysis reveals that BirdMAE, trained on large-scale bird song data with a self-supervised objective, achieves the best performance on the BirdSet benchmark.
  arXiv  Detail & Related papers  (2025-08-02T09:15:16Z)
• Semi-supervised classification of bird vocalizations [0.0]
  Changes in bird populations can indicate broader changes in ecosystems.<n>We propose a semi-supervised acoustic bird detector to allow the detection of time-overlapping calls.<n>It achieves a mean F0.5 score of 0.701 across 315 classes from 110 bird species on a hold-out test set.
  arXiv  Detail & Related papers  (2025-02-19T05:31:13Z)
• Generalization in birdsong classification: impact of transfer learning methods and dataset characteristics [2.6740633963478095]
  We explore the effectiveness of transfer learning in large-scale bird sound classification. Our experiments demonstrate that both fine-tuning and knowledge distillation yield strong performance. We advocate for more comprehensive labeling practices within the animal sound community.
  arXiv  Detail & Related papers  (2024-09-21T11:33:12Z)
• Co-training for Low Resource Scientific Natural Language Inference [65.37685198688538]
  We propose a novel co-training method that assigns weights based on the training dynamics of the classifiers to the distantly supervised labels. By assigning importance weights instead of filtering out examples based on an arbitrary threshold on the predicted confidence, we maximize the usage of automatically labeled data. The proposed method obtains an improvement of 1.5% in Macro F1 over the distant supervision baseline, and substantial improvements over several other strong SSL baselines.
  arXiv  Detail & Related papers  (2024-06-20T18:35:47Z)
• animal2vec and MeerKAT: A self-supervised transformer for rare-event raw audio input and a large-scale reference dataset for bioacoustics [2.1019401515721583]
  animal2vec is an interpretable large transformer model that learns from unlabeled audio and refines its understanding with labeled data. Meerkat Audio Transcripts is the largest labeled dataset on non-human terrestrial mammals. Our model outperforms existing methods on MeerKAT and the publicly available NIPS4Bplus birdsong dataset.
  arXiv  Detail & Related papers  (2024-06-03T12:11:01Z)
• Leveraging tropical reef, bird and unrelated sounds for superior transfer learning in marine bioacoustics [5.799126142155481]
  Machine learning has the potential to revolutionize passive acoustic monitoring (PAM) for ecological assessments. Generalizable pretrained networks can overcome these costs, but high-quality pretraining requires vast annotated libraries. Here, we identify the optimum pretraining strategy for a data-deficient domain using coral reef bioacoustics.
  arXiv  Detail & Related papers  (2024-04-25T09:12:35Z)
• AudioProtoPNet: An interpretable deep learning model for bird sound classification [1.49199020343864]
  This study introduces AudioProtoPNet, an adaptation of the Prototypical Part Network (ProtoPNet) for multi-label bird sound classification. It is an inherently interpretable model that uses a ConvNeXt backbone to extract embeddings. The model was trained on the BirdSet training dataset, which consists of 9,734 bird species and over 6,800 hours of recordings.
  arXiv  Detail & Related papers  (2024-04-16T09:37:41Z)
• WhaleNet: a Novel Deep Learning Architecture for Marine Mammals Vocalizations on Watkins Marine Mammal Sound Database [49.1574468325115]
  We introduce textbfWhaleNet (Wavelet Highly Adaptive Learning Ensemble Network), a sophisticated deep ensemble architecture for the classification of marine mammal vocalizations. We achieve an improvement in classification accuracy by $8-10%$ over existing architectures, corresponding to a classification accuracy of $97.61%$.
  arXiv  Detail & Related papers  (2024-02-20T11:36:23Z)
• TWINS: A Fine-Tuning Framework for Improved Transferability of Adversarial Robustness and Generalization [89.54947228958494]
  This paper focuses on the fine-tuning of an adversarially pre-trained model in various classification tasks. We propose a novel statistics-based approach, Two-WIng NormliSation (TWINS) fine-tuning framework. TWINS is shown to be effective on a wide range of image classification datasets in terms of both generalization and robustness.
  arXiv  Detail & Related papers  (2023-03-20T14:12:55Z)
• Ensembling Off-the-shelf Models for GAN Training [55.34705213104182]
  We find that pretrained computer vision models can significantly improve performance when used in an ensemble of discriminators. We propose an effective selection mechanism, by probing the linear separability between real and fake samples in pretrained model embeddings. Our method can improve GAN training in both limited data and large-scale settings.
  arXiv  Detail & Related papers  (2021-12-16T18:59:50Z)
• Prototypical Classifier for Robust Class-Imbalanced Learning [64.96088324684683]
  We propose textitPrototypical, which does not require fitting additional parameters given the embedding network. Prototypical produces balanced and comparable predictions for all classes even though the training set is class-imbalanced. We test our method on CIFAR-10LT, CIFAR-100LT and Webvision datasets, observing that Prototypical obtains substaintial improvements compared with state of the arts.
  arXiv  Detail & Related papers  (2021-10-22T01:55:01Z)
• Fine-grained Angular Contrastive Learning with Coarse Labels [72.80126601230447]
  We introduce a novel 'Angular normalization' module that allows to effectively combine supervised and self-supervised contrastive pre-training. This work will help to pave the way for future research on this new, challenging, and very practical topic of C2FS classification.
  arXiv  Detail & Related papers  (2020-12-07T08:09:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.