Why the “+”? This year, we’re expanding beyond birds—our train and test data now include mammals, amphibians, and insects as well!
\nBirdCLEF+ offers an exciting opportunity to showcase your skills while contributing to global conservation efforts. This year’s challenge is particularly relevant to conservation projects in Colombia, where climate change and habitat loss continue to pose serious threats to biodiversity.
\nAs your competition hosts, we’re here to support you as much as possible—while ensuring the integrity of the test data remains intact. So, while we’ll be happy to answer questions, we won’t be revealing any secrets about the dataset! 😉
\nNow, let’s introduce the team:
\n@stefankahl – I’m a research associate at the K. Lisa Yang Center for Conservation Bioacoustics (Cornell Lab of Ornithology) and Chemnitz University of Technology. My work focuses on developing machine learning models for detecting and identifying bird species in large-scale audio datasets. I also serve as the technology lead for the BirdNET project and have been organizing the BirdCLEF Challenge since 2018. Feel free to ask me about deep learning for bioacoustics or working note submissions.
@tomdenton – A research software engineer at Google and a long-time co-host of the BirdCLEF Kaggle competitions. His team is dedicated to organizing natural soundscapes in ways that support conservation—essentially, solving core ML challenges so that ecologists can focus on answering scientific questions rather than annotating data and training classifiers. You can check out some of his past work on audio separation for birdsong and source-free domain adaptation. Tom can help with anything related to cutting-edge ML research.
@jscanass – A PhD student at University College London’s People and Nature Lab, working at the intersection of bioacoustics and machine learning. His research focuses on developing datasets and computational models for monitoring bats, amphibians, and insects to support biodiversity studies. Juan is the data collection lead for this competition and can help with questions regarding the Colombian fauna and how it was recorded.
@holgerklinck – Director of the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology. Holger leads efforts in developing and applying innovative conservation technologies to record and analyze sounds in nature at ecologically relevant scales. He can provide insights on bioacoustic monitoring and conservation technology.
@sohier & @maggiemd – Representing the Kaggle team, they’re here to help with competition rules, platform-related issues, and all things Kaggle-specific.
We’re excited to see what you’ll build—happy Kaggling! 🚀
", "messages": [ { "id": 3146298, "postDate": "2025-03-10T18:06:01.350Z", "content": "Why the “+”? This year, we’re expanding beyond birds—our train and test data now include mammals, amphibians, and insects as well!
\nBirdCLEF+ offers an exciting opportunity to showcase your skills while contributing to global conservation efforts. This year’s challenge is particularly relevant to conservation projects in Colombia, where climate change and habitat loss continue to pose serious threats to biodiversity.
\nAs your competition hosts, we’re here to support you as much as possible—while ensuring the integrity of the test data remains intact. So, while we’ll be happy to answer questions, we won’t be revealing any secrets about the dataset! 😉
\nNow, let’s introduce the team:
\n@stefankahl – I’m a research associate at the K. Lisa Yang Center for Conservation Bioacoustics (Cornell Lab of Ornithology) and Chemnitz University of Technology. My work focuses on developing machine learning models for detecting and identifying bird species in large-scale audio datasets. I also serve as the technology lead for the BirdNET project and have been organizing the BirdCLEF Challenge since 2018. Feel free to ask me about deep learning for bioacoustics or working note submissions.
@tomdenton – A research software engineer at Google and a long-time co-host of the BirdCLEF Kaggle competitions. His team is dedicated to organizing natural soundscapes in ways that support conservation—essentially, solving core ML challenges so that ecologists can focus on answering scientific questions rather than annotating data and training classifiers. You can check out some of his past work on audio separation for birdsong and source-free domain adaptation. Tom can help with anything related to cutting-edge ML research.
@jscanass – A PhD student at University College London’s People and Nature Lab, working at the intersection of bioacoustics and machine learning. His research focuses on developing datasets and computational models for monitoring bats, amphibians, and insects to support biodiversity studies. Juan is the data collection lead for this competition and can help with questions regarding the Colombian fauna and how it was recorded.
@holgerklinck – Director of the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology. Holger leads efforts in developing and applying innovative conservation technologies to record and analyze sounds in nature at ecologically relevant scales. He can provide insights on bioacoustic monitoring and conservation technology.
@sohier & @maggiemd – Representing the Kaggle team, they’re here to help with competition rules, platform-related issues, and all things Kaggle-specific.
We’re excited to see what you’ll build—happy Kaggling! 🚀
", "rawMarkdown": "## Welcome to the 2025 BirdCLEF+ Competition! \n\nWhy the “+”? This year, we’re expanding beyond birds—our train and test data now include mammals, amphibians, and insects as well! \n\nBirdCLEF+ offers an exciting opportunity to showcase your skills while contributing to global conservation efforts. This year’s challenge is particularly relevant to conservation projects in Colombia, where climate change and habitat loss continue to pose serious threats to biodiversity. \n\nAs your competition hosts, we’re here to support you as much as possible—while ensuring the integrity of the test data remains intact. So, while we’ll be happy to answer questions, we won’t be revealing any secrets about the dataset! 😉 \n\nNow, let’s introduce the team: \n\n- @stefankahl – I’m a research associate at the [K. Lisa Yang Center for Conservation Bioacoustics](https://www.birds.cornell.edu/ccb/) (Cornell Lab of Ornithology) and [Chemnitz University of Technology](https://www.tu-chemnitz.de/index.html.en). My work focuses on developing machine learning models for detecting and identifying bird species in large-scale audio datasets. I also serve as the technology lead for the [BirdNET project](https://birdnet.cornell.edu/) and have been organizing the BirdCLEF Challenge since 2018. Feel free to ask me about deep learning for bioacoustics or working note submissions. \n\n- @tomdenton – A research software engineer at Google and a long-time co-host of the BirdCLEF Kaggle competitions. His team is dedicated to organizing natural soundscapes in ways that support conservation—essentially, solving core ML challenges so that ecologists can focus on answering scientific questions rather than annotating data and training classifiers. You can check out some of his past work on [audio separation for birdsong](https://ai.googleblog.com/2022/01/separating-birdsong-in-wild-for.html) and [source-free domain adaptation](https://proceedings.mlr.press/v202/boudiaf23a/boudiaf23a.pdf). Tom can help with anything related to cutting-edge ML research. \n\n- @jscanass – A PhD student at University College London’s [People and Nature Lab](https://www.ucl.ac.uk/biosciences/gee/people-and-nature-lab), working at the intersection of bioacoustics and machine learning. His research focuses on developing datasets and computational models for monitoring bats, amphibians, and insects to support biodiversity studies. Juan is the data collection lead for this competition and can help with questions regarding the Colombian fauna and how it was recorded.\n\n- @holgerklinck – Director of the [K. Lisa Yang Center for Conservation Bioacoustics](https://www.birds.cornell.edu/ccb/) at the [Cornell Lab of Ornithology](https://www.birds.cornell.edu/home). Holger leads efforts in developing and applying innovative conservation technologies to record and analyze sounds in nature at ecologically relevant scales. He can provide insights on bioacoustic monitoring and conservation technology. \n\n- @sohier & @maggiemd – Representing the Kaggle team, they’re here to help with competition rules, platform-related issues, and all things Kaggle-specific. \n\nWe’re excited to see what you’ll build—happy Kaggling! 🚀", "votes": 15 }, { "id": 3151456, "postDate": "2025-03-16T17:54:05.613Z", "content": "I've quite enjoyed participating in these competitions in the past, and greatly looking forward to this year!
\nOut of curiosity, what were the research motivations for expanding birdclef to include non-birds? Is it for the purpose of better isolating bird from non-bird vocalizations, or for broadening the scope of your research? If the latter, the dataset this is still quite biased towards bird species identification--why not include even more non-bird audio?
", "rawMarkdown": "I've quite enjoyed participating in these competitions in the past, and greatly looking forward to this year!\n\nOut of curiosity, what were the research motivations for expanding birdclef to include non-birds? Is it for the purpose of better isolating bird from non-bird vocalizations, or for broadening the scope of your research? If the latter, the dataset this is still quite biased towards bird species identification--why not include even more non-bird audio?", "votes": 3, "replies": [ { "id": 3151555, "postDate": "2025-03-16T20:49:04.353Z", "content": "We want to broaden the taxonomic scope. Yet, birds do have a lobby of birdwatchers who share data - other groups like frogs or insects do not have that support and thus, there is less data available (hence the bias). Additionally, annotating insects is very complicated - often, nobody knows which species made the sounds that were recorded. So, this year's competition is a start, there's hopefully more to come in the future. And: It's still BirdCLEF, so yes, lots of birds :)
", "rawMarkdown": "We want to broaden the taxonomic scope. Yet, birds do have a lobby of birdwatchers who share data - other groups like frogs or insects do not have that support and thus, there is less data available (hence the bias). Additionally, annotating insects is very complicated - often, nobody knows which species made the sounds that were recorded. So, this year's competition is a start, there's hopefully more to come in the future. And: It's still BirdCLEF, so yes, lots of birds :)", "votes": 8, "replies": [ { "id": 3201044, "postDate": "2025-05-13T11:21:14.500Z", "rawMarkdown": "", "isDeleted": true }, { "id": 3201049, "postDate": "2025-05-13T11:24:11.143Z", "rawMarkdown": "", "isDeleted": true } ] } ] } ], "comments": [ { "id": 3151456, "author_name": "thacrobatheskis", "author_url": "", "post_date": "2025-03-16T17:54:05.613000", "content": "I've quite enjoyed participating in these competitions in the past, and greatly looking forward to this year!
\nOut of curiosity, what were the research motivations for expanding birdclef to include non-birds? Is it for the purpose of better isolating bird from non-bird vocalizations, or for broadening the scope of your research? If the latter, the dataset this is still quite biased towards bird species identification--why not include even more non-bird audio?
", "votes": 3, "replies": [ { "id": 3151555, "author_name": "Stefan Kahl", "author_url": "", "post_date": "2025-03-16T20:49:04.353000", "content": "We want to broaden the taxonomic scope. Yet, birds do have a lobby of birdwatchers who share data - other groups like frogs or insects do not have that support and thus, there is less data available (hence the bias). Additionally, annotating insects is very complicated - often, nobody knows which species made the sounds that were recorded. So, this year's competition is a start, there's hopefully more to come in the future. And: It's still BirdCLEF, so yes, lots of birds :)
", "votes": 8, "replies": [ { "id": 3201044, "author_name": "", "author_url": "", "post_date": "2025-05-13T11:21:14.500000", "content": "", "votes": 0, "replies": [] }, { "id": 3201049, "author_name": "", "author_url": "", "post_date": "2025-05-13T11:24:11.143000", "content": "", "votes": 0, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "3146298": "## Welcome to the 2025 BirdCLEF+ Competition! \n\nWhy the “+”? This year, we’re expanding beyond birds—our train and test data now include mammals, amphibians, and insects as well! \n\nBirdCLEF+ offers an exciting opportunity to showcase your skills while contributing to global conservation efforts. This year’s challenge is particularly relevant to conservation projects in Colombia, where climate change and habitat loss continue to pose serious threats to biodiversity. \n\nAs your competition hosts, we’re here to support you as much as possible—while ensuring the integrity of the test data remains intact. So, while we’ll be happy to answer questions, we won’t be revealing any secrets about the dataset! 😉 \n\nNow, let’s introduce the team: \n\n- @stefankahl – I’m a research associate at the [K. Lisa Yang Center for Conservation Bioacoustics](https://www.birds.cornell.edu/ccb/) (Cornell Lab of Ornithology) and [Chemnitz University of Technology](https://www.tu-chemnitz.de/index.html.en). My work focuses on developing machine learning models for detecting and identifying bird species in large-scale audio datasets. I also serve as the technology lead for the [BirdNET project](https://birdnet.cornell.edu/) and have been organizing the BirdCLEF Challenge since 2018. Feel free to ask me about deep learning for bioacoustics or working note submissions. \n\n- @tomdenton – A research software engineer at Google and a long-time co-host of the BirdCLEF Kaggle competitions. His team is dedicated to organizing natural soundscapes in ways that support conservation—essentially, solving core ML challenges so that ecologists can focus on answering scientific questions rather than annotating data and training classifiers. You can check out some of his past work on [audio separation for birdsong](https://ai.googleblog.com/2022/01/separating-birdsong-in-wild-for.html) and [source-free domain adaptation](https://proceedings.mlr.press/v202/boudiaf23a/boudiaf23a.pdf). Tom can help with anything related to cutting-edge ML research. \n\n- @jscanass – A PhD student at University College London’s [People and Nature Lab](https://www.ucl.ac.uk/biosciences/gee/people-and-nature-lab), working at the intersection of bioacoustics and machine learning. His research focuses on developing datasets and computational models for monitoring bats, amphibians, and insects to support biodiversity studies. Juan is the data collection lead for this competition and can help with questions regarding the Colombian fauna and how it was recorded.\n\n- @holgerklinck – Director of the [K. Lisa Yang Center for Conservation Bioacoustics](https://www.birds.cornell.edu/ccb/) at the [Cornell Lab of Ornithology](https://www.birds.cornell.edu/home). Holger leads efforts in developing and applying innovative conservation technologies to record and analyze sounds in nature at ecologically relevant scales. He can provide insights on bioacoustic monitoring and conservation technology. \n\n- @sohier & @maggiemd – Representing the Kaggle team, they’re here to help with competition rules, platform-related issues, and all things Kaggle-specific. \n\nWe’re excited to see what you’ll build—happy Kaggling! 🚀", "3151456": "I've quite enjoyed participating in these competitions in the past, and greatly looking forward to this year!\n\nOut of curiosity, what were the research motivations for expanding birdclef to include non-birds? Is it for the purpose of better isolating bird from non-bird vocalizations, or for broadening the scope of your research? If the latter, the dataset this is still quite biased towards bird species identification--why not include even more non-bird audio?" } }