Greetings and welcome to the 2024 BirdCLEF competition!
\nBirdCLEF presents an unparalleled chance for participants to demonstrate their expertise and contribute to global avian conservation efforts. This year's contest holds significant importance for the conservation of bird species in India, as climate change and habitat degradation pose considerable threats to their survival.
\nAs your competition hosts, we aim to be as engaged and helpful as we can, while ensuring the integrity of the test data remains intact. So, please understand we won't be divulging any secrets about it :)
\nAlright, so let's introduce everyone:
\nI am a research associate within the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology and Chemnitz University of Technology, focusing on the development of advanced machine learning models for automatic detection and identification of bird species in large audio collections. Furthermore, I am the technology lead for the BirdNET project and have been organizing the BirdCLEF Challenge since 2018. Feel free to ask me anything related to Deep Learning for bioacoustics or working note submissions.
\n@tomdenton is a research software engineer at Google Brain, and has been co-hosting the BirdCLEF Kaggle competitions for a while now. His group is interested in organizing the sounds of the natural world and making them useful for conservation purposes. Or, put another way, finding ways to solve the basic ML problems such that ecologists can focus on answering ecological questions, instead of annotating data and training classifiers. You can read about some of his prior work on audio separation for birdsong and source free domain adaptation here. Tom can help you with anything related to cutting-edge ML research.
\nVijay Ramesh - @wgbirds - is a post-doc at the K. Lisa Yang Center for Conservation Bioacoustics and his research is centered on exploring the impact of the environment on the ecology and behavior of tropical montane birds. He is specifically interested in the role of climate and habitat in the formation of bird communities in the biodiversity hotspot Western Ghats. His approach is interdisciplinary, blending techniques from bioacoustics, community science, and historical ecology. Previously, his work with acoustic data uncovered how bird species have adapted to ecological restoration. Vijay collected and helped annotate the data for this competition.
\n@holgerklinck is the director of the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology. The mission of the center is to collect and interpret sounds in nature by developing and applying innovative conservation technologies across ecologically-relevant scales to inspire and inform the conservation of wildlife and habitats. Holger can provide you with insights on bioacoustic monitoring in general.
\n@sohier and @maggiemd support the competition from the Kaggle side and are happy to answer questions about rules and Kaggle-specific topics.
\nHappy Kaggling!
", "messages": [ { "id": 2734606, "postDate": "2024-04-04T07:46:12.457Z", "content": "Greetings and welcome to the 2024 BirdCLEF competition!
\nBirdCLEF presents an unparalleled chance for participants to demonstrate their expertise and contribute to global avian conservation efforts. This year's contest holds significant importance for the conservation of bird species in India, as climate change and habitat degradation pose considerable threats to their survival.
\nAs your competition hosts, we aim to be as engaged and helpful as we can, while ensuring the integrity of the test data remains intact. So, please understand we won't be divulging any secrets about it :)
\nAlright, so let's introduce everyone:
\nI am a research associate within the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology and Chemnitz University of Technology, focusing on the development of advanced machine learning models for automatic detection and identification of bird species in large audio collections. Furthermore, I am the technology lead for the BirdNET project and have been organizing the BirdCLEF Challenge since 2018. Feel free to ask me anything related to Deep Learning for bioacoustics or working note submissions.
\n@tomdenton is a research software engineer at Google Brain, and has been co-hosting the BirdCLEF Kaggle competitions for a while now. His group is interested in organizing the sounds of the natural world and making them useful for conservation purposes. Or, put another way, finding ways to solve the basic ML problems such that ecologists can focus on answering ecological questions, instead of annotating data and training classifiers. You can read about some of his prior work on audio separation for birdsong and source free domain adaptation here. Tom can help you with anything related to cutting-edge ML research.
\nVijay Ramesh - @wgbirds - is a post-doc at the K. Lisa Yang Center for Conservation Bioacoustics and his research is centered on exploring the impact of the environment on the ecology and behavior of tropical montane birds. He is specifically interested in the role of climate and habitat in the formation of bird communities in the biodiversity hotspot Western Ghats. His approach is interdisciplinary, blending techniques from bioacoustics, community science, and historical ecology. Previously, his work with acoustic data uncovered how bird species have adapted to ecological restoration. Vijay collected and helped annotate the data for this competition.
\n@holgerklinck is the director of the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology. The mission of the center is to collect and interpret sounds in nature by developing and applying innovative conservation technologies across ecologically-relevant scales to inspire and inform the conservation of wildlife and habitats. Holger can provide you with insights on bioacoustic monitoring in general.
\n@sohier and @maggiemd support the competition from the Kaggle side and are happy to answer questions about rules and Kaggle-specific topics.
\nHappy Kaggling!
", "rawMarkdown": "Greetings and welcome to the 2024 BirdCLEF competition!\n\nBirdCLEF presents an unparalleled chance for participants to demonstrate their expertise and contribute to global avian conservation efforts. This year's contest holds significant importance for the conservation of bird species in India, as climate change and habitat degradation pose considerable threats to their survival.\n\nAs your competition hosts, we aim to be as engaged and helpful as we can, while ensuring the integrity of the test data remains intact. So, please understand we won't be divulging any secrets about it :)\n\nAlright, so let's introduce everyone:\n\nI am a research associate within 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) and [Chemnitz University of Technology](https://www.tu-chemnitz.de/index.html.en), focusing on the development of advanced machine learning models for automatic detection and identification of bird species in large audio collections. Furthermore, I am 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 anything related to Deep Learning for bioacoustics or working note submissions.\n\n@tomdenton is a research software engineer at Google Brain, and has been co-hosting the BirdCLEF Kaggle competitions for a while now. His group is interested in organizing the sounds of the natural world and making them useful for conservation purposes. Or, put another way, finding ways to solve the basic ML problems such that ecologists can focus on answering ecological questions, instead of annotating data and training classifiers. You can read about some of his prior work on [audio separation for birdsong](https://ai.googleblog.com/2022/01/separating-birdsong-in-wild-for.html) and [source free domain adaptation here](https://proceedings.mlr.press/v202/boudiaf23a/boudiaf23a.pdf). Tom can help you with anything related to cutting-edge ML research.\n\nVijay Ramesh - @wgbirds - is a post-doc at the [K. Lisa Yang Center for Conservation Bioacoustics](https://www.birds.cornell.edu/ccb/) and his research is centered on exploring the impact of the environment on the ecology and behavior of tropical montane birds. He is specifically interested in the role of climate and habitat in the formation of bird communities in the biodiversity hotspot Western Ghats. His approach is interdisciplinary, blending techniques from bioacoustics, community science, and historical ecology. Previously, his work with acoustic data uncovered how bird species have adapted to ecological restoration. Vijay collected and helped annotate the data for this competition.\n\n@holgerklinck is the 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). The mission of the center is to collect and interpret sounds in nature by developing and applying innovative conservation technologies across ecologically-relevant scales to inspire and inform the conservation of wildlife and habitats. Holger can provide you with insights on bioacoustic monitoring in general.\n\n@sohier and @maggiemd support the competition from the Kaggle side and are happy to answer questions about rules and Kaggle-specific topics.\n\nHappy Kaggling!", "votes": 12 }, { "id": 2783010, "postDate": "2024-04-29T14:31:53.690Z", "content": "I belive the first line has to have 2024 and not 2023.
\n:-)
True :) Good catch!
", "rawMarkdown": "True :) Good catch!" } ] }, { "id": 2747779, "postDate": "2024-04-12T04:12:36.077Z", "content": "Thanks for organizing such a great competition.
\nI have a question about Evaluations.
\nThe evaluation metric contest a of -averaged ROC-AUC that skips classes which have no positive labels.\nThis suggests that a class may never appear in test_data. In fact, is there a class (in 182 classes) that never appears in test_data? If so, we do not need to train this class.
", "rawMarkdown": "Thanks for organizing such a great competition.\n\nI have a question about Evaluations.\n\n```\nThe evaluation metric for this contest is a version of macro-averaged ROC-AUC that skips classes which have no true positive labels.\n```\n\nThis suggests that a class may **never** appear in test_data. In fact, is there a class (in 182 classes) that never appears in test_data? If so, we do not need to train this class.", "votes": 1, "replies": [ { "id": 2749238, "postDate": "2024-04-12T23:44:45.127Z", "content": "The provided set of classes are plausible classes for the areas the test data was collected in. This is a problem we have in the real world: We have a good idea of the set of species which might be present in the area, but no idea which ones actually appear, or how many.
", "rawMarkdown": "The provided set of classes are plausible classes for the areas the test data was collected in. This is a problem we have in the real world: We have a good idea of the set of species which might be present in the area, but no idea which ones actually appear, or how many.", "votes": 4 } ] }, { "id": 2850002, "postDate": "2024-06-01T21:22:19.360Z", "content": "What a fun contest
", "rawMarkdown": "What a fun contest" }, { "id": 2804320, "postDate": "2024-05-10T01:22:49.433Z", "content": "I have the foolowing message.Please tech me How to proceed the prpject.
\nBefore the test data prediction I load the test data as follows, and next I proceed the
\nprediction code.But I get error. I can't get test data.
\nHow to get the test data? What does \"During the submission re-run\" mean?
\nPlease advise.
test_audio_dir = '../input/birdclef-2024/test_soundscapes/'
\nfile_list = [f.split('.')[0] for f in sorted(os.listdir(test_audio_dir))]
print('Number of test soundscapes:', len(file_list))
\nFileNotFoundError: [Errno 2] No such file or directory: '../input/birdclef-2024/test_soundscapes/readme.ogg'
", "rawMarkdown": "I have the foolowing message.Please tech me How to proceed the prpject.\n\nBefore the test data prediction I load the test data as follows, and next I proceed the\nprediction code.But I get error. I can't get test data.\nHow to get the test data? What does \"During the submission re-run\" mean?\nPlease advise.\n\n# First, load list of audio files by parsing the test_soundscape folder.\n\ntest_audio_dir = '../input/birdclef-2024/test_soundscapes/'\nfile_list = [f.split('.')[0] for f in sorted(os.listdir(test_audio_dir))]\n\n# At the moment, there should only be a single soundscape visible.\n# During the submission re-run, all other hidden soundscapes\n# will be visible too and can be processed by your notebook.\nprint('Number of test soundscapes:', len(file_list))\n\nFileNotFoundError: [Errno 2] No such file or directory: '../input/birdclef-2024/test_soundscapes/readme.ogg'" }, { "id": 2750512, "postDate": "2024-04-13T17:51:03.710Z", "content": "Thank you for your effort; my profile has finally been verified.
", "rawMarkdown": "Thank you for your effort; my profile has finally been verified." }, { "id": 2745069, "postDate": "2024-04-10T10:50:59.807Z", "content": "Dear Host,
\nI hope this message finds you well. I've been attempting to complete the verification process for the BirdCLE2024 competition to submit my identity. However, I've encountered an issue where the camera fails to activate. Could you please assist?
\nThank you for your attention to this matter.
\nWarm regards,
\nRichard
I belive the first line has to have 2024 and not 2023.
\n:-)
True :) Good catch!
", "votes": 0, "replies": [] } ] }, { "id": 2747779, "author_name": "shinmura0", "author_url": "", "post_date": "2024-04-12T04:12:36.077000", "content": "Thanks for organizing such a great competition.
\nI have a question about Evaluations.
\nThe evaluation metric contest a of -averaged ROC-AUC that skips classes which have no positive labels.\nThis suggests that a class may never appear in test_data. In fact, is there a class (in 182 classes) that never appears in test_data? If so, we do not need to train this class.
", "votes": 1, "replies": [ { "id": 2749238, "author_name": "Tom Denton", "author_url": "", "post_date": "2024-04-12T23:44:45.127000", "content": "The provided set of classes are plausible classes for the areas the test data was collected in. This is a problem we have in the real world: We have a good idea of the set of species which might be present in the area, but no idea which ones actually appear, or how many.
", "votes": 4, "replies": [] } ] }, { "id": 2850002, "author_name": "simjim", "author_url": "", "post_date": "2024-06-01T21:22:19.360000", "content": "What a fun contest
", "votes": 0, "replies": [] }, { "id": 2804320, "author_name": "Eiichi tanaka", "author_url": "", "post_date": "2024-05-10T01:22:49.433000", "content": "I have the foolowing message.Please tech me How to proceed the prpject.
\nBefore the test data prediction I load the test data as follows, and next I proceed the
\nprediction code.But I get error. I can't get test data.
\nHow to get the test data? What does \"During the submission re-run\" mean?
\nPlease advise.
test_audio_dir = '../input/birdclef-2024/test_soundscapes/'
\nfile_list = [f.split('.')[0] for f in sorted(os.listdir(test_audio_dir))]
print('Number of test soundscapes:', len(file_list))
\nFileNotFoundError: [Errno 2] No such file or directory: '../input/birdclef-2024/test_soundscapes/readme.ogg'
", "votes": 0, "replies": [] }, { "id": 2750512, "author_name": "Richard", "author_url": "", "post_date": "2024-04-13T17:51:03.710000", "content": "Thank you for your effort; my profile has finally been verified.
", "votes": 0, "replies": [] }, { "id": 2745069, "author_name": "Richard", "author_url": "", "post_date": "2024-04-10T10:50:59.807000", "content": "Dear Host,
\nI hope this message finds you well. I've been attempting to complete the verification process for the BirdCLE2024 competition to submit my identity. However, I've encountered an issue where the camera fails to activate. Could you please assist?
\nThank you for your attention to this matter.
\nWarm regards,
\nRichard