It's wonderful to see how this competition attracts so many people, from the same geographical region, that were not active on Kaggle before!
\nAlso very wonderful that all of these people somehow found each other and are teaming up together.
\nEven more wonderful to see how all of these teams have such similar scores!
\nSurely this cannot be related to the medal selling that have been occurring in many competition lately?
", "messages": [ { "id": "2297148", "postDate": "06/12/2023 11:34:11", "content": "It's wonderful to see how this competition attracts so many people, from the same geographical region, that were not active on Kaggle before!
\nAlso very wonderful that all of these people somehow found each other and are teaming up together.
\nEven more wonderful to see how all of these teams have such similar scores!
\nSurely this cannot be related to the medal selling that have been occurring in many competition lately?
", "rawMarkdown": "It's wonderful to see how this competition attracts so many people, from the same geographical region, that were not active on Kaggle before! \n\nAlso very wonderful that all of these people somehow found each other and are teaming up together. \n\nEven more wonderful to see how all of these teams have such similar scores!\n\nSurely this cannot be related to the medal selling that have been occurring in many competition lately?", "votes": null }, { "id": "2297154", "postDate": "06/12/2023 11:46:28", "content": "I'm sarcasm overdosed =)
", "rawMarkdown": "I'm sarcasm overdosed =)", "votes": null }, { "id": "2297211", "postDate": "06/12/2023 12:22:47", "content": "When the deleted account was deleted 11 days ago, it ranked 19. Now it rank 50. There are dozens of teams with LB higher than 0.4 , whose members are all novices, suddenly occurred after that.
\nI believe that among these teams, there must be some innocent ones. Perhaps some of the honest teams discovered the magic of this competition or maybe some teams are experts in this field themselves so they can get a remarkable score within a week. However, is it possible to have over 30 teams achieving such a significant improvement in the last 11 days?
", "rawMarkdown": "When the deleted account was deleted 11 days ago, it ranked 19. Now it rank 50. There are dozens of teams with LB higher than 0.4 , whose members are all novices, suddenly occurred after that. \n\nI believe that among these teams, there must be some innocent ones. Perhaps some of the honest teams discovered the magic of this competition or maybe some teams are experts in this field themselves so they can get a remarkable score within a week. However, is it possible to have over 30 teams achieving such a significant improvement in the last 11 days?", "votes": null }, { "id": "2297272", "postDate": "06/12/2023 13:24:09", "content": "Well, I have to remind, that it is not a regular Kaggle competition. It is a competition, connected to the Image Matching Challenge at CVPR.
\nIf you check https://www.cs.ubc.ca/research/image-matching-challenge/2021/leaderboard/ - last IMC not on Kaggle, there were many \"noname on Kaggle\" people.
\nhttps://www.youtube.com/live/9cVV9m_b5Ys?feature=share&t=11721
\nE.g. this is the paper from the IMC2021 participants https://twitter.com/ducha_aiki/status/1667135220343099393
My guess, assuming good intentions, would be that many of these people indeed don't care about Kaggle competitions in general, however, they are PhD or Master students, working on image matching. There are many papers (good and bad) from the \"same geographical region\", if you check my twitter.
\nThe same could be true for me - if I hasn't discovered Kaggle long time ago, my first competition related to my PhD studies would probably be this one https://www.kaggle.com/competitions/landmark-retrieval-challenge
\nHowever, I don't have any information to say if any of participant made the (prohibited) private knowledge sharing outside the team. Or maybe formally not prohibited - in some unknown, but open forum or slack channel. We hope that experienced Kaggle team staff would take care of it.
", "rawMarkdown": "Well, I have to remind, that it is not a regular Kaggle competition. It is a competition, connected to the Image Matching Challenge at CVPR. \nIf you check https://www.cs.ubc.ca/research/image-matching-challenge/2021/leaderboard/ - last IMC not on Kaggle, there were many \"noname on Kaggle\" people. \nhttps://www.youtube.com/live/9cVV9m_b5Ys?feature=share&t=11721\nE.g. this is the paper from the IMC2021 participants https://twitter.com/ducha_aiki/status/1667135220343099393\n\n\nMy guess, assuming good intentions, would be that many of these people indeed don't care about Kaggle competitions in general, however, they are PhD or Master students, working on image matching. There are many papers (good and bad) from the \"same geographical region\", if you check my twitter. \n\nThe same could be true for me - if I hasn't discovered Kaggle long time ago, my first competition related to my PhD studies would probably be this one https://www.kaggle.com/competitions/landmark-retrieval-challenge\n\nHowever, I don't have any information to say if any of participant made the (prohibited) private knowledge sharing outside the team. Or maybe formally not prohibited - in some unknown, but open forum or slack channel. We hope that experienced Kaggle team staff would take care of it.", "votes": null }, { "id": "2297284", "postDate": "06/12/2023 13:36:29", "content": "\n\nWell, I have to remind, that it is not a regular Kaggle competition. It is a competition, connected to the Image Matching Challenge at CVPR.
\n
\n If you check https://www.cs.ubc.ca/research/image-matching-challenge/2021/leaderboard/ - last IMC not on Kaggle, there were many \"noname on Kaggle\" people.
\n https://www.youtube.com/live/9cVV9m_b5Ys?feature=share&t=11721
\n E.g. this is the paper from the IMC2021 participants https://twitter.com/ducha_aiki/status/1667135220343099393My guess, assuming good intentions, would be that many of these people indeed don't care about Kaggle competitions in general, however, they are PhD or Master students, working on image matching. There are many papers (good and bad) from the \"same geographical region\", if you check my twitter.
\nThe same could be true for me - if I hasn't discovered Kaggle long time ago, my first competition related to my PhD studies would probably be this one https://www.kaggle.com/competitions/landmark-retrieval-challenge
\nHowever, I don't have any information to say if any of participant made the (prohibited) private knowledge sharing outside the team. Or maybe formally not prohibited - in some unknown, but open forum or slack channel. We hope that experienced Kaggle team staff would take care of it.
\n
Hi @oldufo
\nThank you for your input! I agreed that it is definitely possible that this is coincidence. It just seems to follow a pattern that we have seen in many past competitions in Kaggle before.
\nThe reason of this post is that I do want to bring it to the attention of the people from Kaggle that a more thorough check on private sharing should be done after this competition. Since this is a code competition, this will be rather easy to do!
\nMoreover, maybe could try to find some advertisements of medals being sold within this specific competition?
", "rawMarkdown": "> Well, I have to remind, that it is not a regular Kaggle competition. It is a competition, connected to the Image Matching Challenge at CVPR. \n> If you check https://www.cs.ubc.ca/research/image-matching-challenge/2021/leaderboard/ - last IMC not on Kaggle, there were many \"noname on Kaggle\" people. \n> https://www.youtube.com/live/9cVV9m_b5Ys?feature=share&t=11721\n> E.g. this is the paper from the IMC2021 participants https://twitter.com/ducha_aiki/status/1667135220343099393\n> \n> \n> My guess, assuming good intentions, would be that many of these people indeed don't care about Kaggle competitions in general, however, they are PhD or Master students, working on image matching. There are many papers (good and bad) from the \"same geographical region\", if you check my twitter. \n> \n> The same could be true for me - if I hasn't discovered Kaggle long time ago, my first competition related to my PhD studies would probably be this one https://www.kaggle.com/competitions/landmark-retrieval-challenge\n> \n> However, I don't have any information to say if any of participant made the (prohibited) private knowledge sharing outside the team. Or maybe formally not prohibited - in some unknown, but open forum or slack channel. We hope that experienced Kaggle team staff would take care of it.\n\nHi @oldufo \n\nThank you for your input! I agreed that it is definitely possible that this is coincidence. It just seems to follow a pattern that we have seen in many past competitions in Kaggle before.\n\nThe reason of this post is that I do want to bring it to the attention of the people from Kaggle that a more thorough check on private sharing should be done after this competition. Since this is a code competition, this will be rather easy to do!\n\nMoreover, maybe could try to find some advertisements of medals being sold within this specific competition?", "votes": null }, { "id": "2297289", "postDate": "06/12/2023 13:39:57", "content": "Hi @oldufo , @group16 ， I have found some selling advertisements. Please check my latest post
", "rawMarkdown": "Hi @oldufo , @group16 ， I have found some selling advertisements. Please check my [latest post](https://www.kaggle.com/competitions/image-matching-challenge-2023/discussion/416647)", "votes": null }, { "id": "2297290", "postDate": "06/12/2023 13:39:58", "content": "It seems that someone already found that
\nhttps://www.kaggle.com/competitions/image-matching-challenge-2023/discussion/416647
There are always medal sellers in some social media apps such as Little Red Book, ZhihuThere areThere's always medal sellers in some social media apps such as Little Red Book, ZhiHu. 😐
", "rawMarkdown": "There are always medal sellers in some social media apps such as Little Red Book, ZhihuThere areThere's always medal sellers in some social media apps such as Little Red Book, ZhiHu. 😐", "votes": null }, { "id": "2297629", "postDate": "06/12/2023 17:31:57", "content": "When you see that some top ranked team members are new accounts, it is difficult not to doubt that they have cheating behavior
", "rawMarkdown": "When you see that some top ranked team members are new accounts, it is difficult not to doubt that they have cheating behavior", "votes": null } ], "comments": [ { "id": 2297154, "author_name": "vostankovich", "author_url": "", "post_date": "06/12/2023 11:46:28", "content": "I'm sarcasm overdosed =)
", "votes": null, "replies": [] }, { "id": 2297211, "author_name": "forcewithme", "author_url": "", "post_date": "06/12/2023 12:22:47", "content": "When the deleted account was deleted 11 days ago, it ranked 19. Now it rank 50. There are dozens of teams with LB higher than 0.4 , whose members are all novices, suddenly occurred after that.
\nI believe that among these teams, there must be some innocent ones. Perhaps some of the honest teams discovered the magic of this competition or maybe some teams are experts in this field themselves so they can get a remarkable score within a week. However, is it possible to have over 30 teams achieving such a significant improvement in the last 11 days?
", "votes": null, "replies": [ { "id": 2297310, "author_name": "pamilovedl", "author_url": "", "post_date": "06/12/2023 13:53:47", "content": "There are always medal sellers in some social media apps such as Little Red Book, ZhihuThere areThere's always medal sellers in some social media apps such as Little Red Book, ZhiHu. 😐
", "votes": null, "replies": [] }, { "id": 2297629, "author_name": "bent1e", "author_url": "", "post_date": "06/12/2023 17:31:57", "content": "When you see that some top ranked team members are new accounts, it is difficult not to doubt that they have cheating behavior
", "votes": null, "replies": [] } ] }, { "id": 2297272, "author_name": "oldufo", "author_url": "", "post_date": "06/12/2023 13:24:09", "content": "Well, I have to remind, that it is not a regular Kaggle competition. It is a competition, connected to the Image Matching Challenge at CVPR.
\nIf you check https://www.cs.ubc.ca/research/image-matching-challenge/2021/leaderboard/ - last IMC not on Kaggle, there were many \"noname on Kaggle\" people.
\nhttps://www.youtube.com/live/9cVV9m_b5Ys?feature=share&t=11721
\nE.g. this is the paper from the IMC2021 participants https://twitter.com/ducha_aiki/status/1667135220343099393
My guess, assuming good intentions, would be that many of these people indeed don't care about Kaggle competitions in general, however, they are PhD or Master students, working on image matching. There are many papers (good and bad) from the \"same geographical region\", if you check my twitter.
\nThe same could be true for me - if I hasn't discovered Kaggle long time ago, my first competition related to my PhD studies would probably be this one https://www.kaggle.com/competitions/landmark-retrieval-challenge
\nHowever, I don't have any information to say if any of participant made the (prohibited) private knowledge sharing outside the team. Or maybe formally not prohibited - in some unknown, but open forum or slack channel. We hope that experienced Kaggle team staff would take care of it.
", "votes": null, "replies": [ { "id": 2297284, "author_name": "group16", "author_url": "", "post_date": "06/12/2023 13:36:29", "content": "\n\nWell, I have to remind, that it is not a regular Kaggle competition. It is a competition, connected to the Image Matching Challenge at CVPR.
\n
\n If you check https://www.cs.ubc.ca/research/image-matching-challenge/2021/leaderboard/ - last IMC not on Kaggle, there were many \"noname on Kaggle\" people.
\n https://www.youtube.com/live/9cVV9m_b5Ys?feature=share&t=11721
\n E.g. this is the paper from the IMC2021 participants https://twitter.com/ducha_aiki/status/1667135220343099393My guess, assuming good intentions, would be that many of these people indeed don't care about Kaggle competitions in general, however, they are PhD or Master students, working on image matching. There are many papers (good and bad) from the \"same geographical region\", if you check my twitter.
\nThe same could be true for me - if I hasn't discovered Kaggle long time ago, my first competition related to my PhD studies would probably be this one https://www.kaggle.com/competitions/landmark-retrieval-challenge
\nHowever, I don't have any information to say if any of participant made the (prohibited) private knowledge sharing outside the team. Or maybe formally not prohibited - in some unknown, but open forum or slack channel. We hope that experienced Kaggle team staff would take care of it.
\n
Hi @oldufo
\nThank you for your input! I agreed that it is definitely possible that this is coincidence. It just seems to follow a pattern that we have seen in many past competitions in Kaggle before.
\nThe reason of this post is that I do want to bring it to the attention of the people from Kaggle that a more thorough check on private sharing should be done after this competition. Since this is a code competition, this will be rather easy to do!
\nMoreover, maybe could try to find some advertisements of medals being sold within this specific competition?
", "votes": null, "replies": [ { "id": 2297289, "author_name": "forcewithme", "author_url": "", "post_date": "06/12/2023 13:39:57", "content": "Hi @oldufo , @group16 ， I have found some selling advertisements. Please check my latest post
", "votes": null, "replies": [] }, { "id": 2297290, "author_name": "oldufo", "author_url": "", "post_date": "06/12/2023 13:39:58", "content": "It seems that someone already found that
\nhttps://www.kaggle.com/competitions/image-matching-challenge-2023/discussion/416647