{"id":712,"date":"2024-02-02T08:59:21","date_gmt":"2024-02-02T07:59:21","guid":{"rendered":"https:\/\/websites.fraunhofer.de\/deepbirddetect\/?page_id=712"},"modified":"2025-02-27T14:03:19","modified_gmt":"2025-02-27T13:03:19","slug":"publikationen","status":"publish","type":"page","link":"https:\/\/deepbirddetect.de\/en\/publikationen\/","title":{"rendered":"Publications"},"content":{"rendered":"<div id=\"trailimageid\"><img decoding=\"async\" id=\"ttimg\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/loading.gif\"><\/div>  \n <div class=\"bibsonomycsl_jump_list\">[<a class=\"bibsonomycsl_publications-headline-jumplabel\" href=\"#jmp_2025\" title=\"Goto 2025\">2025<\/a>] [<a class=\"bibsonomycsl_publications-headline-jumplabel\" href=\"#jmp_2024\" title=\"Goto 2024\">2024<\/a>] [<a class=\"bibsonomycsl_publications-headline-jumplabel\" href=\"#jmp_2023\" title=\"Goto 2023\">2023<\/a>]<\/div><ul class=\"bibsonomycsl_publications\">\n<\/ul>\n<a class=\"bibsonomycsl_publications-headline-anchor\" name=\"jmp_2025\"><\/a><h3 class=\"bibsonomycsl_publications-headline\" style=\"font-size: 1.1em; font-weight: bold;\">2025<\/h3>\n<ul class=\"bibsonomycsl_publications\"><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/article.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Brunk, Kristin, H. Kramer, M. Peery, Stefan Kahl, and Connor Wood. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Assessing Spatial Variability and Efficacy of Surrogate Species at an Ecosystem Scale<\/span>\u201d<\/span>. <i>Conservation Biology<\/i> (May 2025). doi:10.1111\/cobi.70058.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-8473c7bacc0696c92c7e549c4fbb8dd5\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-8473c7bacc0696c92c7e549c4fbb8dd5\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/dx.doi.org\/10.1111\/cobi.70058\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-8473c7bacc0696c92c7e549c4fbb8dd5\"><p>@article{article,<br\/>  author = {Brunk, Kristin and Kramer, H. and Peery, M. and Kahl, Stefan and Wood, Connor},<br\/>  journal = {Conservation Biology},<br\/>  keywords = {deepbirddetect},<br\/>  month = {05},<br\/>  title = {Assessing spatial variability and efficacy of surrogate species at an ecosystem scale},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-8473c7bacc0696c92c7e549c4fbb8dd5\"><p>%0 Journal Article<br\/>%1 article<br\/>%A Brunk, Kristin<br\/>%A Kramer, H.<br\/>%A Peery, M.<br\/>%A Kahl, Stefan<br\/>%A Wood, Connor<br\/>%D 2025<br\/>%J Conservation Biology<br\/>%R 10.1111\/cobi.70058<br\/>%T Assessing spatial variability and efficacy of surrogate species at an ecosystem scale<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/article.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Mann, David, Austin Anderson, Amy Donner, Michael Hall, Stefan Kahl, and Holger Klinck. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Continental-Scale Behavioral Response of Birds to a Total Solar Eclipse<\/span>\u201d<\/span>. <i>Scientific Reports<\/i> 15 (April 2025). doi:10.1038\/s41598-025-94901-6.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-227a6a179cf9ec4044156fede2ba7251\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-227a6a179cf9ec4044156fede2ba7251\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/dx.doi.org\/10.1038\/s41598-025-94901-6\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-227a6a179cf9ec4044156fede2ba7251\"><p>@article{article,<br\/>  author = {Mann, David and Anderson, Austin and Donner, Amy and Hall, Michael and Kahl, Stefan and Klinck, Holger},<br\/>  journal = {Scientific Reports},<br\/>  keywords = {deepbirddetect},<br\/>  month = {04},<br\/>  title = {Continental-scale behavioral response of birds to a total solar eclipse},<br\/>  volume = 15,<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-227a6a179cf9ec4044156fede2ba7251\"><p>%0 Journal Article<br\/>%1 article<br\/>%A Mann, David<br\/>%A Anderson, Austin<br\/>%A Donner, Amy<br\/>%A Hall, Michael<br\/>%A Kahl, Stefan<br\/>%A Klinck, Holger<br\/>%D 2025<br\/>%J Scientific Reports<br\/>%R 10.1038\/s41598-025-94901-6<br\/>%T Continental-scale behavioral response of birds to a total solar eclipse<br\/>%V 15<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/misc.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Schwinger, Raphael, Paria Vali Zadeh, Lukas Rauch, Mats Kurz, Tom Hauschild, Sam Lapp, and Sven Tomforde. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Foundation Models for Bioacoustics -- a Comparative Review<\/span>\u201d<\/span>. https:\/\/arxiv.org\/abs\/2508.01277.<\/div>\n<\/div><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/abs\/2508.01277\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-d5791d4ac2c1e9bb6e0ee451fd7a8f18\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-d5791d4ac2c1e9bb6e0ee451fd7a8f18\" href=\"#\">EndNote<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-d5791d4ac2c1e9bb6e0ee451fd7a8f18\"><p>@misc{schwinger2025foundationmodelsbioacoustics,<br\/>  author = {Schwinger, Raphael and Zadeh, Paria Vali and Rauch, Lukas and Kurz, Mats and Hauschild, Tom and Lapp, Sam and Tomforde, Sven},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Foundation Models for Bioacoustics -- a Comparative Review},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-d5791d4ac2c1e9bb6e0ee451fd7a8f18\"><p>%0 Generic<br\/>%1 schwinger2025foundationmodelsbioacoustics<br\/>%A Schwinger, Raphael<br\/>%A Zadeh, Paria Vali<br\/>%A Rauch, Lukas<br\/>%A Kurz, Mats<br\/>%A Hauschild, Tom<br\/>%A Lapp, Sam<br\/>%A Tomforde, Sven<br\/>%D 2025<br\/>%T Foundation Models for Bioacoustics -- a Comparative Review<br\/>%U https:\/\/arxiv.org\/abs\/2508.01277<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/misc.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Heinrich, Ren\u00e9, Lukas Rauch, Bernhard Sick, and Christoph Scholz. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Adversarial Training Improves Generalization Under Distribution Shifts in Bioacoustics<\/span>\u201d<\/span>. https:\/\/arxiv.org\/abs\/2507.13727.<\/div>\n<\/div><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/abs\/2507.13727\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-e014186d962746d1805a04b44dae1f54\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-e014186d962746d1805a04b44dae1f54\" href=\"#\">EndNote<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-e014186d962746d1805a04b44dae1f54\"><p>@misc{heinrich2025adversarialtrainingimprovesgeneralization,<br\/>  author = {Heinrich, Ren\u00e9 and Rauch, Lukas and Sick, Bernhard and Scholz, Christoph},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Adversarial Training Improves Generalization Under Distribution Shifts in Bioacoustics},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-e014186d962746d1805a04b44dae1f54\"><p>%0 Generic<br\/>%1 heinrich2025adversarialtrainingimprovesgeneralization<br\/>%A Heinrich, Ren\u00e9<br\/>%A Rauch, Lukas<br\/>%A Sick, Bernhard<br\/>%A Scholz, Christoph<br\/>%D 2025<br\/>%T Adversarial Training Improves Generalization Under Distribution Shifts in Bioacoustics<br\/>%U https:\/\/arxiv.org\/abs\/2507.13727<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/misc.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Rauch, Lukas, Ren\u00e9 Heinrich, Houtan Ghaffari, Lukas Miklautz, Ilyass Moummad, Bernhard Sick, and Christoph Scholz. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Unmute the Patch Tokens: Rethinking Probing in Multi-Label Audio Classification<\/span>\u201d<\/span>. https:\/\/arxiv.org\/abs\/2509.24901.<\/div>\n<\/div><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/abs\/2509.24901\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-9c7ad0dbfc46b3fae601c4bf7815a170\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-9c7ad0dbfc46b3fae601c4bf7815a170\" href=\"#\">EndNote<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-9c7ad0dbfc46b3fae601c4bf7815a170\"><p>@misc{rauch2025unmutepatchtokensrethinking,<br\/>  author = {Rauch, Lukas and Heinrich, Ren\u00e9 and Ghaffari, Houtan and Miklautz, Lukas and Moummad, Ilyass and Sick, Bernhard and Scholz, Christoph},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Unmute the Patch Tokens: Rethinking Probing in Multi-Label Audio Classification},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-9c7ad0dbfc46b3fae601c4bf7815a170\"><p>%0 Generic<br\/>%1 rauch2025unmutepatchtokensrethinking<br\/>%A Rauch, Lukas<br\/>%A Heinrich, Ren\u00e9<br\/>%A Ghaffari, Houtan<br\/>%A Miklautz, Lukas<br\/>%A Moummad, Ilyass<br\/>%A Sick, Bernhard<br\/>%A Scholz, Christoph<br\/>%D 2025<br\/>%T Unmute the Patch Tokens: Rethinking Probing in Multi-Label Audio Classification<br\/>%U https:\/\/arxiv.org\/abs\/2509.24901<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/misc.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Rauch, Lukas, Ilyass Moummad, Ren\u00e9 Heinrich, Alexis Joly, Bernhard Sick, and Christoph Scholz. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Can Masked Autoencoders Also Listen to Birds?<\/span>\u201d<\/span>. doi:https:\/\/doi.org\/10.48550\/arXiv.2504.12880.<\/div>\n<\/div><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/abs\/2504.12880\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-cb3c659a93748c50b84859ee7ed7ab9b\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-cb3c659a93748c50b84859ee7ed7ab9b\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.48550\/arXiv.2504.12880\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-cb3c659a93748c50b84859ee7ed7ab9b\"><p>@misc{rauch2025maskedautoencoderslistenbirds,<br\/>  author = {Rauch, Lukas and Moummad, Ilyass and Heinrich, Ren\u00e9 and Joly, Alexis and Sick, Bernhard and Scholz, Christoph},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Can Masked Autoencoders Also Listen to Birds?},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-cb3c659a93748c50b84859ee7ed7ab9b\"><p>%0 Generic<br\/>%1 rauch2025maskedautoencoderslistenbirds<br\/>%A Rauch, Lukas<br\/>%A Moummad, Ilyass<br\/>%A Heinrich, Ren\u00e9<br\/>%A Joly, Alexis<br\/>%A Sick, Bernhard<br\/>%A Scholz, Christoph<br\/>%D 2025<br\/>%R https:\/\/doi.org\/10.48550\/arXiv.2504.12880<br\/>%T Can Masked Autoencoders Also Listen to Birds?<br\/>%U https:\/\/arxiv.org\/abs\/2504.12880<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/article.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Weidlich-Rau, Melissa, Amanda K. Navine, Patrick T. Chaopricha, Felix G\u00fcnther, Stefan Kahl, Thomas Wilhelm-Stein, Raymond C. Mack, et al. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Continuous Real-Time Acoustic Monitoring of Endangered Bird Species in Hawai\u2018i<\/span>\u201d<\/span>. <i>Ecological Informatics<\/i> 87 (2025): 103102. doi:https:\/\/doi.org\/10.1016\/j.ecoinf.2025.103102.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-449f5830d7768ef66e59f3d6f439e669\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/www.sciencedirect.com\/science\/article\/pii\/S1574954125001116\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-449f5830d7768ef66e59f3d6f439e669\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-449f5830d7768ef66e59f3d6f439e669\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.1016\/j.ecoinf.2025.103102\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-449f5830d7768ef66e59f3d6f439e669\">The decline of endemic bird species in Hawai\u2018i requires innovative conservation measures enabled by passive acoustic monitoring (PAM). This paper describes a novel real-time PAM system used in the P\u014dhakuloa Training Area (PTA) to reduce wildlife collisions and minimize disruptions to military operations while ensuring the protection of endangered bird species such as the N\u0113n\u0113 and \u2018Ak\u0113\u2018ak\u0113. The system is based on the BirdNET algorithm and was evaluated with over 16,000 soundscape recordings from Hawai\u2018i. The results show that the model version HI V2.0, based on BirdNET and specifically adapted to Hawaiian bird species, showed the clearest separation between true and false positive detections (average precision 49% to 52%), although this difference was not statistically significant. However, accuracy varied considerably between species and locations, emphasizing the need to adapt the models to the specific conditions of use. A novel web application allows immediate visualization of the predicted bird species, facilitating the implementation of conservation measures. The three acoustic monitoring units installed at the PTA in January 2023 demonstrate the system\u2019s potential for continuous monitoring and protection of Hawaiian endangered bird species.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-449f5830d7768ef66e59f3d6f439e669\"><p>@article{WEIDLICHRAU2025103102,<br\/>  abstract = {The decline of endemic bird species in Hawai\u2018i requires innovative conservation measures enabled by passive acoustic monitoring (PAM). This paper describes a novel real-time PAM system used in the P\u014dhakuloa Training Area (PTA) to reduce wildlife collisions and minimize disruptions to military operations while ensuring the protection of endangered bird species such as the N\u0113n\u0113 and \u2018Ak\u0113\u2018ak\u0113. The system is based on the BirdNET algorithm and was evaluated with over 16,000 soundscape recordings from Hawai\u2018i. The results show that the model version HI V2.0, based on BirdNET and specifically adapted to Hawaiian bird species, showed the clearest separation between true and false positive detections (average precision 49% to 52%), although this difference was not statistically significant. However, accuracy varied considerably between species and locations, emphasizing the need to adapt the models to the specific conditions of use. A novel web application allows immediate visualization of the predicted bird species, facilitating the implementation of conservation measures. The three acoustic monitoring units installed at the PTA in January 2023 demonstrate the system\u2019s potential for continuous monitoring and protection of Hawaiian endangered bird species.},<br\/>  author = {Weidlich-Rau, Melissa and Navine, Amanda K. and Chaopricha, Patrick T. and G\u00fcnther, Felix and Kahl, Stefan and Wilhelm-Stein, Thomas and Mack, Raymond C. and Reers, Hendrik and Rice, Aaron N. and Eibl, Maximilian and Hart, Patrick J. and Wolff, Patrick and Klinck, Holger and Schnell, Lena D. and Doratt, Rogelio and Loquet, Michael and Lackey, Tiana},<br\/>  journal = {Ecological Informatics},<br\/>  keywords = {deepbirddetect},<br\/>  pages = 103102,<br\/>  title = {Continuous Real-Time Acoustic Monitoring of endangered bird species in Hawai\u2018i},<br\/>  volume = 87,<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-449f5830d7768ef66e59f3d6f439e669\"><p>%0 Journal Article<br\/>%1 WEIDLICHRAU2025103102<br\/>%A Weidlich-Rau, Melissa<br\/>%A Navine, Amanda K.<br\/>%A Chaopricha, Patrick T.<br\/>%A G\u00fcnther, Felix<br\/>%A Kahl, Stefan<br\/>%A Wilhelm-Stein, Thomas<br\/>%A Mack, Raymond C.<br\/>%A Reers, Hendrik<br\/>%A Rice, Aaron N.<br\/>%A Eibl, Maximilian<br\/>%A Hart, Patrick J.<br\/>%A Wolff, Patrick<br\/>%A Klinck, Holger<br\/>%A Schnell, Lena D.<br\/>%A Doratt, Rogelio<br\/>%A Loquet, Michael<br\/>%A Lackey, Tiana<br\/>%D 2025<br\/>%J Ecological Informatics<br\/>%P 103102<br\/>%R https:\/\/doi.org\/10.1016\/j.ecoinf.2025.103102<br\/>%T Continuous Real-Time Acoustic Monitoring of endangered bird species in Hawai\u2018i<br\/>%U https:\/\/www.sciencedirect.com\/science\/article\/pii\/S1574954125001116<br\/>%V 87<br\/>%X The decline of endemic bird species in Hawai\u2018i requires innovative conservation measures enabled by passive acoustic monitoring (PAM). This paper describes a novel real-time PAM system used in the P\u014dhakuloa Training Area (PTA) to reduce wildlife collisions and minimize disruptions to military operations while ensuring the protection of endangered bird species such as the N\u0113n\u0113 and \u2018Ak\u0113\u2018ak\u0113. The system is based on the BirdNET algorithm and was evaluated with over 16,000 soundscape recordings from Hawai\u2018i. The results show that the model version HI V2.0, based on BirdNET and specifically adapted to Hawaiian bird species, showed the clearest separation between true and false positive detections (average precision 49% to 52%), although this difference was not statistically significant. However, accuracy varied considerably between species and locations, emphasizing the need to adapt the models to the specific conditions of use. A novel web application allows immediate visualization of the predicted bird species, facilitating the implementation of conservation measures. The three acoustic monitoring units installed at the PTA in January 2023 demonstrate the system\u2019s potential for continuous monitoring and protection of Hawaiian endangered bird species.<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/inproceedings.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Rauch, Lukas, Raphael Schwinger, Moritz Wirth, Ren\u00e9 Heinrich, Denis Huseljic, Marek Herde, Jonas Lange, et al. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">BirdSet: A Large-Scale Dataset for Audio Classification in Avian Bioacoustics<\/span>\u201d<\/span>. In <i>International Conference on Learning Representations (ICLR)<\/i>. ICLR, 2025. https:\/\/arxiv.org\/abs\/2403.10380v6.<\/div>\n<\/div><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/abs\/2403.10380v6\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-ea77eb240131466adfef529c35f3ea62\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-ea77eb240131466adfef529c35f3ea62\" href=\"#\">EndNote<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-ea77eb240131466adfef529c35f3ea62\"><p>@inproceedings{rauch2024birdset,<br\/>  author = {Rauch, Lukas and Schwinger, Raphael and Wirth, Moritz and Heinrich, Ren\u00e9 and Huseljic, Denis and Herde, Marek and Lange, Jonas and Kahl, Stefan and Sick, Bernhard and Tomforde, Sven and Scholz, Christoph},<br\/>  booktitle = {International Conference on Learning Representations (ICLR)},<br\/>  keywords = {deepbirddetect},<br\/>  publisher = {ICLR},<br\/>  title = {BirdSet: A Large-Scale Dataset for Audio Classification in Avian Bioacoustics},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-ea77eb240131466adfef529c35f3ea62\"><p>%0 Conference Paper<br\/>%1 rauch2024birdset<br\/>%A Rauch, Lukas<br\/>%A Schwinger, Raphael<br\/>%A Wirth, Moritz<br\/>%A Heinrich, Ren\u00e9<br\/>%A Huseljic, Denis<br\/>%A Herde, Marek<br\/>%A Lange, Jonas<br\/>%A Kahl, Stefan<br\/>%A Sick, Bernhard<br\/>%A Tomforde, Sven<br\/>%A Scholz, Christoph<br\/>%B International Conference on Learning Representations (ICLR)<br\/>%D 2025<br\/>%I ICLR<br\/>%T BirdSet: A Large-Scale Dataset for Audio Classification in Avian Bioacoustics<br\/>%U https:\/\/arxiv.org\/abs\/2403.10380v6<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/article.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Heinrich, Ren\u00e9, Lukas Rauch, Bernhard Sick, and Christoph Scholz. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">AudioProtoPNet: An Interpretable Deep Learning Model for Bird Sound Classification<\/span>\u201d<\/span>. <i>Ecological Informatics<\/i> (2025): 103081. doi:https:\/\/doi.org\/10.1016\/j.ecoinf.2025.103081.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-e35720bf678cc3d8a205409980742aee\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/www.sciencedirect.com\/science\/article\/pii\/S1574954125000901\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-e35720bf678cc3d8a205409980742aee\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-e35720bf678cc3d8a205409980742aee\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.1016\/j.ecoinf.2025.103081\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-e35720bf678cc3d8a205409980742aee\">Deep learning models have significantly advanced acoustic bird monitoring by recognizing numerous bird species based on their vocalizations. However, traditional deep learning models are black boxes that provide no insight into their underlying computations, limiting their usefulness to ornithologists and machine learning engineers. Explainable models could facilitate debugging, knowledge discovery, trust, and interdisciplinary collaboration. We introduce AudioProtoPNet, an adaptation of the Prototypical Part Network (ProtoPNet) for multi-label bird sound classification. It is inherently interpretable, leveraging a ConvNeXt backbone to extract embeddings and a prototype learning classifier trained on these embeddings. The classifier learns prototypical patterns of each bird species\u2019 vocalizations from spectrograms of instances in the training data. During inference, recordings are classified by comparing them to learned prototypes in the embedding space, providing explanations for the model\u2019s decisions and insights into the most informative embeddings of each bird species. The model was trained on the BirdSet training dataset, which consists of 9734 bird species and over 6800 h of recordings. Its performance was evaluated on the seven BirdSet test datasets, covering different geographical regions. AudioProtoPNet outperformed the state-of-the-art bird sound classification model Perch, which is superior to the more popular BirdNet, achieving an average AUROC of 0.90 and a cmAP of 0.42, with relative improvements of 7.1% and 16.7% over Perch, respectively. These results demonstrate that even for the challenging task of multi-label bird sound classification, it is possible to develop powerful yet interpretable deep learning models that provide valuable insights for professionals in ornithology and machine learning.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-e35720bf678cc3d8a205409980742aee\"><p>@article{HEINRICH2025103081,<br\/>  abstract = {Deep learning models have significantly advanced acoustic bird monitoring by recognizing numerous bird species based on their vocalizations. However, traditional deep learning models are black boxes that provide no insight into their underlying computations, limiting their usefulness to ornithologists and machine learning engineers. Explainable models could facilitate debugging, knowledge discovery, trust, and interdisciplinary collaboration. We introduce AudioProtoPNet, an adaptation of the Prototypical Part Network (ProtoPNet) for multi-label bird sound classification. It is inherently interpretable, leveraging a ConvNeXt backbone to extract embeddings and a prototype learning classifier trained on these embeddings. The classifier learns prototypical patterns of each bird species\u2019 vocalizations from spectrograms of instances in the training data. During inference, recordings are classified by comparing them to learned prototypes in the embedding space, providing explanations for the model\u2019s decisions and insights into the most informative embeddings of each bird species. The model was trained on the BirdSet training dataset, which consists of 9734 bird species and over 6800 h of recordings. Its performance was evaluated on the seven BirdSet test datasets, covering different geographical regions. AudioProtoPNet outperformed the state-of-the-art bird sound classification model Perch, which is superior to the more popular BirdNet, achieving an average AUROC of 0.90 and a cmAP of 0.42, with relative improvements of 7.1% and 16.7% over Perch, respectively. These results demonstrate that even for the challenging task of multi-label bird sound classification, it is possible to develop powerful yet interpretable deep learning models that provide valuable insights for professionals in ornithology and machine learning.},<br\/>  author = {Heinrich, Ren\u00e9 and Rauch, Lukas and Sick, Bernhard and Scholz, Christoph},<br\/>  journal = {Ecological Informatics},<br\/>  keywords = 2025,<br\/>  pages = 103081,<br\/>  title = {AudioProtoPNet: An interpretable deep learning model for bird sound classification},<br\/>  year = 2025<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-e35720bf678cc3d8a205409980742aee\"><p>%0 Journal Article<br\/>%1 HEINRICH2025103081<br\/>%A Heinrich, Ren\u00e9<br\/>%A Rauch, Lukas<br\/>%A Sick, Bernhard<br\/>%A Scholz, Christoph<br\/>%D 2025<br\/>%J Ecological Informatics<br\/>%P 103081<br\/>%R https:\/\/doi.org\/10.1016\/j.ecoinf.2025.103081<br\/>%T AudioProtoPNet: An interpretable deep learning model for bird sound classification<br\/>%U https:\/\/www.sciencedirect.com\/science\/article\/pii\/S1574954125000901<br\/>%X Deep learning models have significantly advanced acoustic bird monitoring by recognizing numerous bird species based on their vocalizations. However, traditional deep learning models are black boxes that provide no insight into their underlying computations, limiting their usefulness to ornithologists and machine learning engineers. Explainable models could facilitate debugging, knowledge discovery, trust, and interdisciplinary collaboration. We introduce AudioProtoPNet, an adaptation of the Prototypical Part Network (ProtoPNet) for multi-label bird sound classification. It is inherently interpretable, leveraging a ConvNeXt backbone to extract embeddings and a prototype learning classifier trained on these embeddings. The classifier learns prototypical patterns of each bird species\u2019 vocalizations from spectrograms of instances in the training data. During inference, recordings are classified by comparing them to learned prototypes in the embedding space, providing explanations for the model\u2019s decisions and insights into the most informative embeddings of each bird species. The model was trained on the BirdSet training dataset, which consists of 9734 bird species and over 6800 h of recordings. Its performance was evaluated on the seven BirdSet test datasets, covering different geographical regions. AudioProtoPNet outperformed the state-of-the-art bird sound classification model Perch, which is superior to the more popular BirdNet, achieving an average AUROC of 0.90 and a cmAP of 0.42, with relative improvements of 7.1% and 16.7% over Perch, respectively. These results demonstrate that even for the challenging task of multi-label bird sound classification, it is possible to develop powerful yet interpretable deep learning models that provide valuable insights for professionals in ornithology and machine learning.<br\/><\/p><\/div><\/div><\/li>\n<\/ul>\n<a class=\"bibsonomycsl_publications-headline-anchor\" name=\"jmp_2024\"><\/a><h3 class=\"bibsonomycsl_publications-headline\" style=\"font-size: 1.1em; font-weight: bold;\">2024<\/h3>\n<ul class=\"bibsonomycsl_publications\"><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/article.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Wood, Connor M., Felix G\u00fcnther, Angela Rex, Daniel F. Hofstadter, Hendrik Reers, Stefan Kahl, M. Zachariah Peery, and Holger Klinck. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Real-Time Acoustic Monitoring Facilitates the Proactive Management of Biological Invasions<\/span>\u201d<\/span>. <i>Biological Invasions<\/i> 26, no. 12 (December 1, 2024): 3989\u20133996. doi:10.1007\/s10530-024-03426-y.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-acd8a4cfe249b5e2a0577d328817bb5e\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.1007\/s10530-024-03426-y\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-acd8a4cfe249b5e2a0577d328817bb5e\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-acd8a4cfe249b5e2a0577d328817bb5e\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/dx.doi.org\/10.1007\/s10530-024-03426-y\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-acd8a4cfe249b5e2a0577d328817bb5e\">Biological surveillance at an invasion front is hindered by low population densities and, among animals, high mobility of target species. Using the barred owl (Strix varia) invasion of western North American forests as a test case, we tested real-time autonomous recording units (the ecoPi, OekoFor GbR, Freiburg, Germany) by deploying them in an area known to be occupied by the target species. The ecoPi passively record audio, analyze it onboard with the BirdNET algorithm, and transmit audio clips with identifiable sounds via cellular network to a web interface where users can listen to audio to manually vet the results. We successfully detected and lethally removed three barred owls, demonstrating that real-time acoustic monitoring can be used to support rapid interventions at the forefront of an ongoing invasion in which proactive management may be essential to the protection of an iconic native species, the spotted owl (S. occidentalis). This approach has the potential to make a significant contribution to global biodiversity conservation efforts by massively increasing the speed at which biological invasions by acoustically active species, and other time-sensitive conservation challenges, can be managed.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-acd8a4cfe249b5e2a0577d328817bb5e\"><p>@article{Wood2024,<br\/>  abstract = {Biological surveillance at an invasion front is hindered by low population densities and, among animals, high mobility of target species. Using the barred owl (Strix varia) invasion of western North American forests as a test case, we tested real-time autonomous recording units (the ecoPi, OekoFor GbR, Freiburg, Germany) by deploying them in an area known to be occupied by the target species. The ecoPi passively record audio, analyze it onboard with the BirdNET algorithm, and transmit audio clips with identifiable sounds via cellular network to a web interface where users can listen to audio to manually vet the results. We successfully detected and lethally removed three barred owls, demonstrating that real-time acoustic monitoring can be used to support rapid interventions at the forefront of an ongoing invasion in which proactive management may be essential to the protection of an iconic native species, the spotted owl (S. occidentalis). This approach has the potential to make a significant contribution to global biodiversity conservation efforts by massively increasing the speed at which biological invasions by acoustically active species, and other time-sensitive conservation challenges, can be managed.},<br\/>  author = {Wood, Connor M. and G\u00fcnther, Felix and Rex, Angela and Hofstadter, Daniel F. and Reers, Hendrik and Kahl, Stefan and Peery, M. Zachariah and Klinck, Holger},<br\/>  journal = {Biological Invasions},<br\/>  keywords = {deepbirddetect},<br\/>  month = 12,<br\/>  number = 12,<br\/>  pages = {3989--3996},<br\/>  title = {Real-time acoustic monitoring facilitates the proactive management of biological invasions},<br\/>  volume = 26,<br\/>  year = 2024<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-acd8a4cfe249b5e2a0577d328817bb5e\"><p>%0 Journal Article<br\/>%1 Wood2024<br\/>%A Wood, Connor M.<br\/>%A G\u00fcnther, Felix<br\/>%A Rex, Angela<br\/>%A Hofstadter, Daniel F.<br\/>%A Reers, Hendrik<br\/>%A Kahl, Stefan<br\/>%A Peery, M. Zachariah<br\/>%A Klinck, Holger<br\/>%D 2024<br\/>%J Biological Invasions<br\/>%N 12<br\/>%P 3989--3996<br\/>%R 10.1007\/s10530-024-03426-y<br\/>%T Real-time acoustic monitoring facilitates the proactive management of biological invasions<br\/>%U https:\/\/doi.org\/10.1007\/s10530-024-03426-y<br\/>%V 26<br\/>%X Biological surveillance at an invasion front is hindered by low population densities and, among animals, high mobility of target species. Using the barred owl (Strix varia) invasion of western North American forests as a test case, we tested real-time autonomous recording units (the ecoPi, OekoFor GbR, Freiburg, Germany) by deploying them in an area known to be occupied by the target species. The ecoPi passively record audio, analyze it onboard with the BirdNET algorithm, and transmit audio clips with identifiable sounds via cellular network to a web interface where users can listen to audio to manually vet the results. We successfully detected and lethally removed three barred owls, demonstrating that real-time acoustic monitoring can be used to support rapid interventions at the forefront of an ongoing invasion in which proactive management may be essential to the protection of an iconic native species, the spotted owl (S. occidentalis). This approach has the potential to make a significant contribution to global biodiversity conservation efforts by massively increasing the speed at which biological invasions by acoustically active species, and other time-sensitive conservation challenges, can be managed.<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border bibsonomycsl_preview_thumb\">\n                                        <span>\n                                            <img decoding=\"async\" class=\"bibsonomycsl_preview\" style=\"z-index: 1;\" src=\"https:\/\/deepbirddetect.de\/wp-content\/plugins\/bibsonomy-csl\/img\/entrytypes\/article.jpg\" \/>\n                                        <\/span>\n                                 <\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Rauch, Lukas, Denis Huseljic, Moritz Wirth, Jens Decke, Bernhard Sick, and Christoph Scholz. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Towards Deep Active Learning in Avian Bioacoustics.<\/span>\u201d<\/span>. <i>CoRR<\/i> abs\/2406.18621 (2024). http:\/\/dblp.uni-trier.de\/db\/journals\/corr\/corr2406.html#abs-2406-18621.<\/div>\n<\/div><span class=\"bibsonomycsl_url\"><a href=\"http:\/\/dblp.uni-trier.de\/db\/journals\/corr\/corr2406.html#abs-2406-18621\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-109c7a8537fd559cc852b1c9272d117d\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-109c7a8537fd559cc852b1c9272d117d\" href=\"#\">EndNote<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-109c7a8537fd559cc852b1c9272d117d\"><p>@article{journals\/corr\/abs-2406-18621,<br\/>  author = {Rauch, Lukas and Huseljic, Denis and Wirth, Moritz and Decke, Jens and Sick, Bernhard and Scholz, Christoph},<br\/>  journal = {CoRR},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Towards Deep Active Learning in Avian Bioacoustics.},<br\/>  volume = {abs\/2406.18621},<br\/>  year = 2024<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-109c7a8537fd559cc852b1c9272d117d\"><p>%0 Journal Article<br\/>%1 journals\/corr\/abs-2406-18621<br\/>%A Rauch, Lukas<br\/>%A Huseljic, Denis<br\/>%A Wirth, Moritz<br\/>%A Decke, Jens<br\/>%A Sick, Bernhard<br\/>%A Scholz, Christoph<br\/>%D 2024<br\/>%J CoRR<br\/>%T Towards Deep Active Learning in Avian Bioacoustics.<br\/>%U http:\/\/dblp.uni-trier.de\/db\/journals\/corr\/corr2406.html#abs-2406-18621<br\/>%V abs\/2406.18621<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border\"><img decoding=\"async\" onmouseover=\"javascript:showtrail('https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=07a18630671c300494c159f15d4fcf69&fileName=s10336-024-02144-5.pdf&size=LARGE')\" onmouseout=\"javascript:hidetrail()\" class=\"bibsonomycsl_preview\" src=\"https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=07a18630671c300494c159f15d4fcf69&fileName=s10336-024-02144-5.pdf&size=SMALL&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000stream%5D=Resource id #49&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000seekable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000readable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000writable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000uri%5D=php:\/\/temp&\" \/><\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Wood, Connor M., and Stefan Kahl. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Guidelines for Approriate Use of BirdNET Scores and Other Detector Outputs<\/span>\u201d<\/span>. <i>Journal of Ornithology<\/i> (2024). doi:https:\/\/doi.org\/10.1007\/s10336-024-02144-5.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-07a18630671c300494c159f15d4fcf69\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.1007\/s10336-024-02144-5\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-07a18630671c300494c159f15d4fcf69\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-07a18630671c300494c159f15d4fcf69\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.1007\/s10336-024-02144-5\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-07a18630671c300494c159f15d4fcf69\">Machine learning tools capable of identifying animals by sound have proliferated, making the challenge of interpreting their outputs much more prevalent. These tools, like their predecessors, quantify prediction uncertainty with scores that tend to resemble probabilities but are actually unitless scores that are (generally) positively related to prediction accuracy in species-specific ways. BirdNET is one such tool, a freely available animal sound identification algorithm capable of identifying\u2009>\u20096,000 species, most of them birds. We describe two ways in which BirdNET \u201cconfidence scores\u201d\u2014and the output scores of other detector tools\u2014can be used appropriately to interpret BirdNET results (reviewing them down to a user-defined threshold or converting them to probabilities), and provide a step-by-step tutorial to follow these suggestions. These suggestions are complementary to common performance metrics like precision, recall, and receiver operating characteristic. BirdNET can be a powerful tool for acoustic-based biodiversity research, but its utility depends on the careful use and interpretation of its outputs.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-07a18630671c300494c159f15d4fcf69\"><p>@article{wood2024guidelines,<br\/>  abstract = {Machine learning tools capable of identifying animals by sound have proliferated, making the challenge of interpreting their outputs much more prevalent. These tools, like their predecessors, quantify prediction uncertainty with scores that tend to resemble probabilities but are actually unitless scores that are (generally) positively related to prediction accuracy in species-specific ways. BirdNET is one such tool, a freely available animal sound identification algorithm capable of identifying\u2009>\u20096,000 species, most of them birds. We describe two ways in which BirdNET \u201cconfidence scores\u201d\u2014and the output scores of other detector tools\u2014can be used appropriately to interpret BirdNET results (reviewing them down to a user-defined threshold or converting them to probabilities), and provide a step-by-step tutorial to follow these suggestions. These suggestions are complementary to common performance metrics like precision, recall, and receiver operating characteristic. BirdNET can be a powerful tool for acoustic-based biodiversity research, but its utility depends on the careful use and interpretation of its outputs.},<br\/>  author = {Wood, Connor M. and Kahl, Stefan},<br\/>  journal = {Journal of Ornithology},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Guidelines for approriate use of BirdNET scores and other detector outputs},<br\/>  year = 2024<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-07a18630671c300494c159f15d4fcf69\"><p>%0 Journal Article<br\/>%1 wood2024guidelines<br\/>%A Wood, Connor M.<br\/>%A Kahl, Stefan<br\/>%D 2024<br\/>%J Journal of Ornithology<br\/>%R https:\/\/doi.org\/10.1007\/s10336-024-02144-5<br\/>%T Guidelines for approriate use of BirdNET scores and other detector outputs<br\/>%U https:\/\/doi.org\/10.1007\/s10336-024-02144-5<br\/>%X Machine learning tools capable of identifying animals by sound have proliferated, making the challenge of interpreting their outputs much more prevalent. These tools, like their predecessors, quantify prediction uncertainty with scores that tend to resemble probabilities but are actually unitless scores that are (generally) positively related to prediction accuracy in species-specific ways. BirdNET is one such tool, a freely available animal sound identification algorithm capable of identifying\u2009>\u20096,000 species, most of them birds. We describe two ways in which BirdNET \u201cconfidence scores\u201d\u2014and the output scores of other detector tools\u2014can be used appropriately to interpret BirdNET results (reviewing them down to a user-defined threshold or converting them to probabilities), and provide a step-by-step tutorial to follow these suggestions. These suggestions are complementary to common performance metrics like precision, recall, and receiver operating characteristic. BirdNET can be a powerful tool for acoustic-based biodiversity research, but its utility depends on the careful use and interpretation of its outputs.<br\/><\/p><\/div><\/div><\/li>\n<\/ul>\n<a class=\"bibsonomycsl_publications-headline-anchor\" name=\"jmp_2023\"><\/a><h3 class=\"bibsonomycsl_publications-headline\" style=\"font-size: 1.1em; font-weight: bold;\">2023<\/h3>\n<ul class=\"bibsonomycsl_publications\"><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border\"><img decoding=\"async\" onmouseover=\"javascript:showtrail('https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=ab59c9114d872e33a3ce3e101453e888&fileName=s41598-023-49989-z.pdf&size=LARGE')\" onmouseout=\"javascript:hidetrail()\" class=\"bibsonomycsl_preview\" src=\"https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=ab59c9114d872e33a3ce3e101453e888&fileName=s41598-023-49989-z.pdf&size=SMALL&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000stream%5D=Resource id #54&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000seekable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000readable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000writable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000uri%5D=php:\/\/temp&\" \/><\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Ghani, Burooj, Tom Denton, Stefan Kahl, and Holger Klinck. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Global Birdsong Embeddings Enable Superior Transfer Learning for Bioacoustic Classification<\/span>\u201d<\/span>. <i>Scientific Reports<\/i> (2023). doi:doi:10.1038\/s41598-023-49989-z.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-ab59c9114d872e33a3ce3e101453e888\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-ab59c9114d872e33a3ce3e101453e888\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-ab59c9114d872e33a3ce3e101453e888\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/dx.doi.org\/doi:10.1038\/s41598-023-49989-z\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-ab59c9114d872e33a3ce3e101453e888\">Automated bioacoustic analysis aids understanding and protection of both marine and terrestrial animals and their habitats across extensive spatiotemporal scales, and typically involves analyzing vast collections of acoustic data. With the advent of deep learning models, classification of important signals from these datasets has markedly improved. These models power critical data analyses for research and decision-making in biodiversity monitoring, animal behaviour studies, and natural resource management. However, deep learning models are often data-hungry and require a significant amount of labeled training data to perform well. While sufficient training data is available for certain taxonomic groups (e.g., common bird species), many classes (such as rare and endangered species, many non-bird taxa, and call-type) lack enough data to train a robust model from scratch. This study investigates the utility of feature embeddings extracted from audio classification models to identify bioacoustic classes other than the ones these models were originally trained on. We evaluate models on diverse datasets, including different bird calls and dialect types, bat calls, marine mammals calls, and amphibians calls. The embeddings extracted from the models trained on bird vocalization data consistently allowed higher quality classification than the embeddings trained on general audio datasets. The results of this study indicate that high-quality feature embeddings from large-scale acoustic bird classifiers can be harnessed for few-shot transfer learning, enabling the learning of new classes from a limited quantity of training data. Our findings reveal the potential for efficient analyses of novel bioacoustic tasks, even in scenarios where available training data is limited to a few samples.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-ab59c9114d872e33a3ce3e101453e888\"><p>@article{ghani2023global,<br\/>  abstract = {Automated bioacoustic analysis aids understanding and protection of both marine and terrestrial animals and their habitats across extensive spatiotemporal scales, and typically involves analyzing vast collections of acoustic data. With the advent of deep learning models, classification of important signals from these datasets has markedly improved. These models power critical data analyses for research and decision-making in biodiversity monitoring, animal behaviour studies, and natural resource management. However, deep learning models are often data-hungry and require a significant amount of labeled training data to perform well. While sufficient training data is available for certain taxonomic groups (e.g., common bird species), many classes (such as rare and endangered species, many non-bird taxa, and call-type) lack enough data to train a robust model from scratch. This study investigates the utility of feature embeddings extracted from audio classification models to identify bioacoustic classes other than the ones these models were originally trained on. We evaluate models on diverse datasets, including different bird calls and dialect types, bat calls, marine mammals calls, and amphibians calls. The embeddings extracted from the models trained on bird vocalization data consistently allowed higher quality classification than the embeddings trained on general audio datasets. The results of this study indicate that high-quality feature embeddings from large-scale acoustic bird classifiers can be harnessed for few-shot transfer learning, enabling the learning of new classes from a limited quantity of training data. Our findings reveal the potential for efficient analyses of novel bioacoustic tasks, even in scenarios where available training data is limited to a few samples.},<br\/>  author = {Ghani, Burooj and Denton, Tom and Kahl, Stefan and Klinck, Holger},<br\/>  journal = {Scientific Reports},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Global Birdsong Embeddings Enable Superior Transfer Learning for Bioacoustic Classification},<br\/>  year = 2023<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-ab59c9114d872e33a3ce3e101453e888\"><p>%0 Journal Article<br\/>%1 ghani2023global<br\/>%A Ghani, Burooj<br\/>%A Denton, Tom<br\/>%A Kahl, Stefan<br\/>%A Klinck, Holger<br\/>%D 2023<br\/>%J Scientific Reports<br\/>%R doi:10.1038\/s41598-023-49989-z<br\/>%T Global Birdsong Embeddings Enable Superior Transfer Learning for Bioacoustic Classification<br\/>%X Automated bioacoustic analysis aids understanding and protection of both marine and terrestrial animals and their habitats across extensive spatiotemporal scales, and typically involves analyzing vast collections of acoustic data. With the advent of deep learning models, classification of important signals from these datasets has markedly improved. These models power critical data analyses for research and decision-making in biodiversity monitoring, animal behaviour studies, and natural resource management. However, deep learning models are often data-hungry and require a significant amount of labeled training data to perform well. While sufficient training data is available for certain taxonomic groups (e.g., common bird species), many classes (such as rare and endangered species, many non-bird taxa, and call-type) lack enough data to train a robust model from scratch. This study investigates the utility of feature embeddings extracted from audio classification models to identify bioacoustic classes other than the ones these models were originally trained on. We evaluate models on diverse datasets, including different bird calls and dialect types, bat calls, marine mammals calls, and amphibians calls. The embeddings extracted from the models trained on bird vocalization data consistently allowed higher quality classification than the embeddings trained on general audio datasets. The results of this study indicate that high-quality feature embeddings from large-scale acoustic bird classifiers can be harnessed for few-shot transfer learning, enabling the learning of new classes from a limited quantity of training data. Our findings reveal the potential for efficient analyses of novel bioacoustic tasks, even in scenarios where available training data is limited to a few samples.<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border\"><img decoding=\"async\" onmouseover=\"javascript:showtrail('https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=cd5bd23bd249fbf486013f74458703e5&fileName=2308.07121v2.pdf&size=LARGE')\" onmouseout=\"javascript:hidetrail()\" class=\"bibsonomycsl_preview\" src=\"https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=cd5bd23bd249fbf486013f74458703e5&fileName=2308.07121v2.pdf&size=SMALL&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000stream%5D=Resource id #59&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000seekable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000readable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000writable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000uri%5D=php:\/\/temp&\" \/><\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Rauch, Lukas, Raphael Schwinger, Moritz Wirth, Bernhard Sick, Sven Tomforde, and Christoph Scholz. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">Active Bird2Vec: Towards End-To-End Bird Sound Monitoring With Transformers<\/span>\u201d<\/span> (2023). doi:doi:10.48550\/ARXIV.2308.07121.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-cd5bd23bd249fbf486013f74458703e5\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/pdf\/2308.07121v2.pdf\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-cd5bd23bd249fbf486013f74458703e5\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-cd5bd23bd249fbf486013f74458703e5\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/dx.doi.org\/doi:10.48550\/ARXIV.2308.07121\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-cd5bd23bd249fbf486013f74458703e5\">We propose a shift towards end-to-end learning in bird sound monitoring by combining self-supervised (SSL) and deep active learning (DAL). Leveraging transformer models, we aim to bypass traditional spectrogram conversions, enabling direct raw audio processing. ACTIVE BIRD2VEC is set to generate high-quality bird sound representations through SSL, potentially accelerating the assessment of environmental changes and decision-making processes for wind farms. Additionally, we seek to utilize the wide variety of bird vocalizations through DAL, reducing the reliance on extensively labeled datasets by human experts. We plan to curate a comprehensive set of tasks through Huggingface Datasets, enhancing future comparability and reproducibility of bioacoustic research. A comparative analysis between various transformer models will be conducted to evaluate their proficiency in bird sound recognition tasks. We aim to accelerate the progression of avian bioacoustic research and contribute to more effective conservation strategies.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-cd5bd23bd249fbf486013f74458703e5\"><p>@article{rauch2023active,<br\/>  abstract = {We propose a shift towards end-to-end learning in bird sound monitoring by combining self-supervised (SSL) and deep active learning (DAL). Leveraging transformer models, we aim to bypass traditional spectrogram conversions, enabling direct raw audio processing. ACTIVE BIRD2VEC is set to generate high-quality bird sound representations through SSL, potentially accelerating the assessment of environmental changes and decision-making processes for wind farms. Additionally, we seek to utilize the wide variety of bird vocalizations through DAL, reducing the reliance on extensively labeled datasets by human experts. We plan to curate a comprehensive set of tasks through Huggingface Datasets, enhancing future comparability and reproducibility of bioacoustic research. A comparative analysis between various transformer models will be conducted to evaluate their proficiency in bird sound recognition tasks. We aim to accelerate the progression of avian bioacoustic research and contribute to more effective conservation strategies.},<br\/>  author = {Rauch, Lukas and Schwinger, Raphael and Wirth, Moritz and Sick, Bernhard and Tomforde, Sven and Scholz, Christoph},<br\/>  keywords = {deepbirddetect},<br\/>  title = {Active Bird2Vec: Towards End-To-End Bird Sound Monitoring with Transformers},<br\/>  year = 2023<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-cd5bd23bd249fbf486013f74458703e5\"><p>%0 Journal Article<br\/>%1 rauch2023active<br\/>%A Rauch, Lukas<br\/>%A Schwinger, Raphael<br\/>%A Wirth, Moritz<br\/>%A Sick, Bernhard<br\/>%A Tomforde, Sven<br\/>%A Scholz, Christoph<br\/>%D 2023<br\/>%R doi:10.48550\/ARXIV.2308.07121<br\/>%T Active Bird2Vec: Towards End-To-End Bird Sound Monitoring with Transformers<br\/>%U https:\/\/arxiv.org\/pdf\/2308.07121v2.pdf<br\/>%X We propose a shift towards end-to-end learning in bird sound monitoring by combining self-supervised (SSL) and deep active learning (DAL). Leveraging transformer models, we aim to bypass traditional spectrogram conversions, enabling direct raw audio processing. ACTIVE BIRD2VEC is set to generate high-quality bird sound representations through SSL, potentially accelerating the assessment of environmental changes and decision-making processes for wind farms. Additionally, we seek to utilize the wide variety of bird vocalizations through DAL, reducing the reliance on extensively labeled datasets by human experts. We plan to curate a comprehensive set of tasks through Huggingface Datasets, enhancing future comparability and reproducibility of bioacoustic research. A comparative analysis between various transformer models will be conducted to evaluate their proficiency in bird sound recognition tasks. We aim to accelerate the progression of avian bioacoustic research and contribute to more effective conservation strategies.<br\/><\/p><\/div><\/div><\/li><li class=\"bibsonomycsl_pubitem\"><div class=\"bibsonomycsl_preview_border\"><img decoding=\"async\" onmouseover=\"javascript:showtrail('https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=16f5bddcbb89781342d36956c24d1951&fileName=2312.07439.pdf&size=LARGE')\" onmouseout=\"javascript:hidetrail()\" class=\"bibsonomycsl_preview\" src=\"https:\/\/deepbirddetect.de\/en\/publikationen\/?action=preview&userName=dbd&intraHash=16f5bddcbb89781342d36956c24d1951&fileName=2312.07439.pdf&size=SMALL&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000stream%5D=Resource id #64&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000seekable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000readable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000writable%5D=1&doc%5B\u0000GuzzleHttp\\Psr7\\Stream\u0000uri%5D=php:\/\/temp&\" \/><\/div><div class=\"bibsonomycsl_entry\"><div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Hamer, Jenny, Eleni Triantafillou, Bart van Merri\u00ebnboer, Stefan Kahl, Holger Klinck, Tom Denton, and Vincent Dumoulin. <span class=\"csl-title\">\u201c<span class=\"\u2018csl-title\u2019\">BIRB: A Generalization Benchmark for Information Retrieval in Bioacoustics<\/span>\u201d<\/span> (2023). doi:https:\/\/doi.org\/10.48550\/arXiv.2312.07439.<\/div>\n<\/div><span class=\"bibsonomycsl_export bibsonomycsl_abstract\"><a rel=\"abs-16f5bddcbb89781342d36956c24d1951\"  href=\"#\">Abstract<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/arxiv.org\/abs\/2312.07439\" target=\"_blank\">URL<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_bibtex\"><a rel=\"bib-16f5bddcbb89781342d36956c24d1951\" href=\"#\">BibTeX<\/a><\/span><span class=\"bibsonomycsl_export bibsonomycsl_endnote\"><a rel=\"end-16f5bddcbb89781342d36956c24d1951\" href=\"#\">EndNote<\/a><\/span><span class=\"bibsonomycsl_url\"><a href=\"https:\/\/doi.org\/10.48550\/arXiv.2312.07439\" target=\"_blank\">DOI<\/a><\/span><div style=\"clear: left\"> <\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_abstract\" style=\"display:none;\" id=\"abs-16f5bddcbb89781342d36956c24d1951\">The ability for a machine learning model to cope with differences in training and deployment conditions--e.g. in the presence of distribution shift or the generalization to new classes altogether--is crucial for real-world use cases. However, most empirical work in this area has focused on the image domain with artificial benchmarks constructed to measure individual aspects of generalization. We present BIRB, a complex benchmark centered on the retrieval of bird vocalizations from passively-recorded datasets given focal recordings from a large citizen science corpus available for training. We propose a baseline system for this collection of tasks using representation learning and a nearest-centroid search. Our thorough empirical evaluation and analysis surfaces open research directions, suggesting that BIRB fills the need for a more realistic and complex benchmark to drive progress on robustness to distribution shifts and generalization of ML models.<\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_bibtex\" style=\"display:none;\" id=\"bib-16f5bddcbb89781342d36956c24d1951\"><p>@article{hamer2023generalization,<br\/>  abstract = {The ability for a machine learning model to cope with differences in training and deployment conditions--e.g. in the presence of distribution shift or the generalization to new classes altogether--is crucial for real-world use cases. However, most empirical work in this area has focused on the image domain with artificial benchmarks constructed to measure individual aspects of generalization. We present BIRB, a complex benchmark centered on the retrieval of bird vocalizations from passively-recorded datasets given focal recordings from a large citizen science corpus available for training. We propose a baseline system for this collection of tasks using representation learning and a nearest-centroid search. Our thorough empirical evaluation and analysis surfaces open research directions, suggesting that BIRB fills the need for a more realistic and complex benchmark to drive progress on robustness to distribution shifts and generalization of ML models.},<br\/>  author = {Hamer, Jenny and Triantafillou, Eleni and van Merri\u00ebnboer, Bart and Kahl, Stefan and Klinck, Holger and Denton, Tom and Dumoulin, Vincent},<br\/>  keywords = {deepbirddetect},<br\/>  title = {BIRB: A Generalization Benchmark for Information Retrieval in Bioacoustics},<br\/>  year = 2023<br\/>}<br\/><\/p><\/div><div class=\"bibsonomycsl_collapse bibsonomycsl_pub_endnote\" style=\"display:none;\" id=\"end-16f5bddcbb89781342d36956c24d1951\"><p>%0 Journal Article<br\/>%1 hamer2023generalization<br\/>%A Hamer, Jenny<br\/>%A Triantafillou, Eleni<br\/>%A van Merri\u00ebnboer, Bart<br\/>%A Kahl, Stefan<br\/>%A Klinck, Holger<br\/>%A Denton, Tom<br\/>%A Dumoulin, Vincent<br\/>%D 2023<br\/>%R https:\/\/doi.org\/10.48550\/arXiv.2312.07439<br\/>%T BIRB: A Generalization Benchmark for Information Retrieval in Bioacoustics<br\/>%U https:\/\/arxiv.org\/abs\/2312.07439<br\/>%X The ability for a machine learning model to cope with differences in training and deployment conditions--e.g. in the presence of distribution shift or the generalization to new classes altogether--is crucial for real-world use cases. However, most empirical work in this area has focused on the image domain with artificial benchmarks constructed to measure individual aspects of generalization. We present BIRB, a complex benchmark centered on the retrieval of bird vocalizations from passively-recorded datasets given focal recordings from a large citizen science corpus available for training. We propose a baseline system for this collection of tasks using representation learning and a nearest-centroid search. Our thorough empirical evaluation and analysis surfaces open research directions, suggesting that BIRB fills the need for a more realistic and complex benchmark to drive progress on robustness to distribution shifts and generalization of ML models.<br\/><\/p><\/div><\/div><\/li><\/ul>","protected":false},"excerpt":{"rendered":"<p>[2025] [2024] [2023] 2025 Brunk, Kristin, H. Kramer, M. Peery, Stefan Kahl, and Connor Wood. \u201cAssessing Spatial Variability and Efficacy of Surrogate Species at an Ecosystem Scale\u201d. Conservation Biology (May 2025). doi:10.1111\/cobi.70058. BibTeXEndNoteDOI @article{article, author = {Brunk, Kristin and Kramer, H. and Peery, M. and Kahl, Stefan and Wood, Connor}, journal = {Conservation Biology}, keywords [&hellip;]<\/p>","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"class_list":["post-712","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/pages\/712","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/comments?post=712"}],"version-history":[{"count":3,"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/pages\/712\/revisions"}],"predecessor-version":[{"id":736,"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/pages\/712\/revisions\/736"}],"wp:attachment":[{"href":"https:\/\/deepbirddetect.de\/en\/wp-json\/wp\/v2\/media?parent=712"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}