Multispecies bioacoustics classification using transfer learning of deep convolutional neural networks with pseudo-labeling
Abstract
In this study, we evaluated deep convolutional neural networks for classifying the calls of 24 birds and amphibian species detected in ambient field recordings from the tropical mountains of Puerto Rico. Training data were collected using a template-based detection algorithm and manually validated with a graphical interface. To reduce the labor intensive and time-consuming process of manual validation, as well as to increase the accuracy of species classification with acoustic recordings, we propose a novel approach that combines transfer learning of a pre-trained deep convolutional neural network (CNN) model, a semi-supervised pseudo-labeling method, and a custom training loss function. While generating sufficient training data is a major challenge for many deep learning applications, our proposed methodology enables the network to be trained in a supervised fashion with labeled and unlabeled data simultaneously, which effectively increases the size of training set and thus boosts the model performance. As a result, the model achieves 97.7% sensitivity, 96.4% specificity and 96.6% accuracy in classifying a test set of manually validated true and false positive template-based detections. This multi-label multi-species classification methodology and its framework can be easily expanded to other acoustic classification problems.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Semi-supervised learning with pseudo-labeling approaches has achieved >97% sensitivity and >96% specificity across diverse species sets.
“Semi-supervised learning is also closing the annotation gap, with pseudo-labeling approaches hitting > 97% sensitivity and > 96% specificity across diverse species sets”
From Listening forward: emerging roles of bioacoustics in ecology, evolution, and conservation (2026) · cites this paper as Zhong et al. 2020
Cites
Still being checked.
Cited by
Sorted by citations
Exploring the relationship between the soundscape and the environment: A systematic review
Turlington, K., Suárez‐Castro, A. F., Teixeira, D., Linke, S., & Sheldon, F. (2024) · Ecological Indicators · doi:10.1016/j.ecolind.2024.112388 · Cited by 43 · see citing works
Lv, S., Xia, C., Cheng, C., Yan, J., & Wu, X. (2024) · Computers & Industrial Engineering · doi:10.1016/j.cie.2024.110542 · Cited by 13 · see citing works
Windy events detection in big bioacoustics datasets using a pre-trained Convolutional Neural Network
Terranova, F., Betti, L., Ferrario, V., Friard, O., Ludynia, K., Petersen, G. S., Mathevon, N., Reby, D., & Favaro, L. (2024) · The Science of The Total Environment · doi:10.1016/j.scitotenv.2024.174868 · Cited by 11 · see citing works
Correlation-based feature partition regression method for unsupervised anomaly detection
Liu, Z., Gao, X., Jia, X., Xue, B., Fu, S., Li, K., Huang, X., & Huang, Z. (2022) · Applied Intelligence · doi:10.1007/s10489-022-03247-3 · Cited by 6 · see citing works
Syunkova, A., Lapp, S., Basanta, M. D., Lambertini, C., Guzman, S. R., Voyles, J., Richards‐Zawacki, C. L., & Kitzes, J. A. (2025) · Ecological Informatics · doi:10.1016/j.ecoinf.2025.103427 · Cited by 1 · see citing works
Listening forward: emerging roles of bioacoustics in ecology, evolution, and conservation
Zsebők, S. (2026) · Biologia Futura · doi:10.1007/s42977-026-00322-5
“Semi-supervised learning is also closing the annotation gap, with pseudo-labeling approaches hitting > 97% sensitivity and > 96% specificity across diverse species sets (Zhong et al. 2020).”
- Authors
- Zhong, M., LeBien, J., Campos-Cerqueira, M., Dodhia, R., Lavista Ferres, J., Velev, J., & Aide, T. M.
- Affiliations
- Microsoft (United States) (US)
- Sieve Analytics, San Juan
- Microsoft, Redmond, WA
- Corresponding author
- Zhong, Ming
- Venue
- The Journal of the Acoustical Society of America
- Publisher
- Acoustical Society of America (ASA)
- Published
- 2020-10-01
- Volume
- 148
- Issue
- 4_Supplement
- Pages
- 2442-2442
- ISSN
- 0001-4966 (ISSN-L)
- 1520-8524
- OpenAlex
- W3109653336
- Cited by
- 6
- References
- 0
- Citations by year
- 2026: 1 · 2025: 1 · 2024: 3 · 2022: 1
- Type
- conference abstract
- Language
- en
- Open access
- closed
- Locations
- The Journal of the Acoustical Society of America (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Signal Processing › Music and Audio Processing
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Developmental Biology › Animal Vocal Communication and Behavior
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Keywords
- bird call classification, transfer learning, pseudo-labeling, semi-supervised learning, deep convolutional neural networks, template-based detection, acoustic species identification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
