[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82871-en":3,"doc-seo-82871-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82871,8796095462418,"Noah","https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780",8,"Research & Report","Adaptive Diversity-Uncertainty Active Learning with Redundancy Control for Bioacoustic Event Classification","Active learning reduces annotation costs in large-scale bioacoustic monitoring, where expert labeling is costly and audio data varies strongly across environments. Existing selection strategies often rely on static criteria and cannot adapt to changing prediction reliability during training. The report presents an active learning approach for multilabel bioacoustic event classification that jointly leverages predictive uncertainty, embedding-space diversity, and intra-batch redundancy control. An adaptive weighting shifts from exploration to exploitation as confidence grows, while greedy MMR enforces batch diversity. Experiments in BioDCASE 2026 Task 4 improve learning efficiency and achieve competitive macro mAP and AULC across terrestrial and marine benchmarks.","ADAPTIVE DIVERSITY-UNCERTAINTY ACTIVE LEARNING WITH REDUNDANCY CONTROL FOR BIOACOUSTIC EVENT CLASSIFICATION  \nTechnical Report  \nGabriel Dubus 1∗, Hugo Magaldi 1, Anatole Gros-Martial2,  \n1 Eco-Anthropologie, Musum National d’Histoire Naturelle, UMR7206, CNRS, Paris, France,  \n2 Centre d’Etudes Biologiques de Chiz, UMR 7372, CNRS, La Rochelle Universit, France,∗ Corresponding author: Gabriel Dubus: [gabriel.dubus@mnhn.fr](gabriel.dubus@mnhn.fr)  \narXiv :2607 .04868v 1 [ cs . SD] 6 Jul 2026  \nABSTRACT  \nActive learning is a promising framework for reducing annotation costs in large-scale bioacoustic monitoring, where expert labeling is expensive and data distributions are highly heterogeneous across environments. However, existing sample selection strategies often rely on static criteria that do not adapt to the evolving reliability of model predictions during training. This limitation can lead to suboptimal exploration–exploitation trade-offs and redundant sample selection.  \nWe propose an active learning strategy for multilabel bioacoustic event classification that jointly models predictive uncertainty, embedding-space diversity, and intra-batch redundancy. The method introduces an adaptive weighting scheme that progressively shifts from diversity-driven exploration in high-uncertainty regimes toward uncertainty-driven exploitation as the model becomes more confident, reflecting the increasing reliability of the classifier. To further improve annotation efficiency, a greedy Maximum Marginal Relevance (MMR) procedure is used to enforce diversity among selected samples within each acquisition batch.  \nWe evaluate the proposed approach within the BioDCASE 2026 Task 4 active learning framework on terrestrial (BirdSet) and marine (ATBFL) benchmarks using pretrained audio embeddings anda fixed annotation budget. Experimental results show consistent improvements in learning efficiency and competitive in terms of macro mean Average Precision (mAP) and Area Under the Learning Curve (AULC) across heterogeneous acoustic domains. The gains are particularly pronounced on structured terrestrial soundscapes, while performance remains competitive under noisier marine conditions. These findings demonstrate that adaptive acquisition strategies combining uncertainty estimation, embedding-space diversity, and redundancy-aware batch construction provide an effective and robust solution for annotation-efficient bioacoustic learning.  \nIndex Terms— Active Learning, Bioacoustics, Multilabel Classification, Sample Selection, Uncertainty Sampling, Diversity Sampling  \n1. INTRODUCTION  \nLarge-scale bioacoustic monitoring relies on passive acoustic systems that continuously generate massive volumes of unlabeled audio data in both terrestrial and marine environments. Although these recordings contain rich ecological information, only a small fraction can be manually annotated due to the high cost and expertise required for reliable labeling [1] . These annotations are  \ntypically required to train machine learning models for acoustic event detection and classification.  \nThis imbalance between data availability and annotation budget raises a central question: given a large pool of unlabeled acoustic data, how should samples be selected for annotation in order to maximize model performance under a strict labeling budget? This work was submitted to the BioDCASE 2026 Task 4 workshop challenge on Active Learning for Bioacoustics, which specifically addresses annotation-efficient learning across terrestrial and marine acoustic domains. The challenge provides a standardized active learning framework, pretrained PerchV2 embeddings, and fixed classifier training procedures, requiring participants to focus exclusively on the design of acquisition strategies under a constrained annotation budget.  \nActive Learning (AL) provides a principled framework to address this challenge by iteratively selecting the most informative samples for labeling [2] . AL methods aim t","cbCaimHY5GmH6gHS","https://ap.wps.com/l/cbCaimHY5GmH6gHS","pdf",211214,1,5,"English","en",105,"# Abstract\n# Introduction\n## Problem and challenge in bioacoustic annotation\n## BioDCASE Task 4 evaluation setting\n## Proposed unified acquisition strategy and contributions","[{\"question\":\"Why do existing bioacoustic sample selection strategies underperform during training?\",\"answer\":\"They often use static selection criteria and do not adapt to the model’s evolving prediction reliability. This can produce poor exploration–exploitation trade-offs and redundant sample choices.\"},{\"question\":\"What signals does the proposed method use to select samples for annotation?\",\"answer\":\"It combines predictive uncertainty, embedding-space diversity, and intra-batch redundancy control for multilabel bioacoustic event classification.\"},{\"question\":\"How is diversity enforced within each acquisition batch?\",\"answer\":\"A greedy Maximum Marginal Relevance (MMR) procedure selects samples to maintain diversity while reducing redundancy inside each acquisition batch.\"}]",1784183574,13,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"adaptive-diversity-uncertainty-active-learning-with-redundancy-control-for-bioacoustic-event-classification","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/adaptive-diversity-uncertainty-active-learning-with-redundancy-control-for-bioacoustic-event-classification/82871/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why do existing bioacoustic sample selection strategies underperform during training?","Question",{"text":75,"@type":76},"They often use static selection criteria and do not adapt to the model’s evolving prediction reliability. This can produce poor exploration–exploitation trade-offs and redundant sample choices.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What signals does the proposed method use to select samples for annotation?",{"text":80,"@type":76},"It combines predictive uncertainty, embedding-space diversity, and intra-batch redundancy control for multilabel bioacoustic event classification.",{"name":82,"@type":73,"acceptedAnswer":83},"How is diversity enforced within each acquisition batch?",{"text":84,"@type":76},"A greedy Maximum Marginal Relevance (MMR) procedure selects samples to maintain diversity while reducing redundancy inside each acquisition batch.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,109,114,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":21,"slug":137},19,"General","general"]