A finely annotated dataset for the automated acoustic identification of European Orthoptera and Cicadidae

2026-04-07 · Source: Machine learning : nature.com subject feeds · Field: Science & Research — Life Sciences & Biology, Environmental Science & Earth Systems, Artificial Intelligence & Machine Learning · Depth: Advanced, long

Summary

A new dataset, ECOSoundSet, has been released to advance automated acoustic identification of European Orthoptera and Cicadidae, addressing the urgent need for large-scale insect biodiversity monitoring. Published on April 7, 2026, this dataset comprises 11,224 recordings covering 193 orthopteran and 24 cicada species from North, Central, and temperate Western Europe. It integrates both weakly labeled recordings, indicating species presence, and strongly labeled recordings, which precisely specify the time and frequency range of each insect sound. This resource, deposited in a Zenodo repository (DOI: 10.5281/zenodo.15043892), complements existing online biodiversity data and supports the development of robust algorithms for insect sound classification, crucial for tracking widespread declines in insect populations.

Key takeaway

For AI Scientists and Machine Learning Engineers developing biodiversity monitoring tools, this new ECOSoundSet dataset offers a critical resource. Its combination of weak and strong acoustic labels for 217 European insect species can significantly improve the robustness and accuracy of your automated recognition algorithms. You should integrate this dataset into your model training and validation workflows to enhance species identification capabilities, thereby contributing to more effective large-scale insect population monitoring efforts.

Key insights

A new dataset combines weak and strong acoustic labels for automated European insect identification.

Principles

• Passive acoustic monitoring scales insect biodiversity tracking.
• Diverse, heterogeneous datasets improve recognition algorithms.

Method

The dataset combines coarsely labeled recordings (species presence inferred) with finely annotated recordings (time and frequency range specified) to support automated acoustic classification.

In practice

• Access ECOSoundSet via Zenodo DOI: 10.5281/zenodo.15043892.
• Utilize provided scripts for dataset analysis and recording retrieval.
• Evaluate the Hugging Face model for audio segment duration effects.

Topics

• Automated Acoustic Identification
• European Orthoptera
• European Cicadidae
• Passive Acoustic Monitoring
• Weak and Strong Labeling

Code references

• DavidFunosas/ECOSoundSet_database_analysis
• DavidFunosas/GBIF_recording_download

Best for: AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• A fully automated framework for acoustic identification and localization of terrestrial wildlife at scale — Automated acoustic localization using low-cost hardware and open-source software enables large-scale wildlife monitoring.
• Robustness and Diversity Evaluation on ProsSegue-ML: a Free Prosodic Segmentation Tool for Brazilian Portuguese — ProsSegue-ML's robustness and bias are evaluated for Brazilian Portuguese prosodic segmentation.
• A multi-modal dataset for insect biodiversity with imagery and DNA at the trap and individual level — MassID45 integrates bulk and individual insect imagery with DNA data to train high-throughput taxonomic classifiers.
• GetNetUPAM: Ecologically Informed Nested Cross-Validation and Noise-Robust Attention for Marine Bioacoustic Monitoring — Ecologically informed nested cross-validation and adaptive attention improve marine bioacoustic model stability and generalization.
• Dolph2Vec: Self-Supervised Representations of Dolphin Vocalizations — Dolph2Vec, a species-specific SSL model, effectively uncovers fine-grained dolphin communication patterns using a novel, longitudinal dataset.
• How We Used Google's Perch v2 to Build a Bird-Based H5N1 Early Warning System in Patagonia — Bioacoustics foundation models can enable large-scale environmental monitoring for critical conservation challenges.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.