![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F18102%2Ff51edb7061baf429ac2c0d9c74f39c34%2FScreenshot%202020-06-14%20at%2013.09.05.png?generation=1592133038916261&alt=media\")
Given what happened with the team All Faces are Real, I feel like we have a lot of things unanswered in this competition:-
\n\nI am thankful to Kaggle team, as I learnt a lot about image recognition and CV overall, but in a competition with the stakes so high, why did there have to be so much confusion?
\n\nThe Kaggle team acts an intermediary after all - I feel that the hosts have the power regarding rules et cetera so I do not wish to proverbially shoot the messenger. So my question is to Mozaic and the other hosts - could you have made this a better experience for all of us? Would it be possible to?
", "messages": [ { "id": "884808", "postDate": "06/13/2020 16:30:14", "content": "Given what happened with the team All Faces are Real, I feel like we have a lot of things unanswered in this competition:-
\n\nI am thankful to Kaggle team, as I learnt a lot about image recognition and CV overall, but in a competition with the stakes so high, why did there have to be so much confusion?
\n\nThe Kaggle team acts an intermediary after all - I feel that the hosts have the power regarding rules et cetera so I do not wish to proverbially shoot the messenger. So my question is to Mozaic and the other hosts - could you have made this a better experience for all of us? Would it be possible to?
", "rawMarkdown": "Given what happened with the team All Faces are Real, I feel like we have a lot of things unanswered in this competition:-\n\n+ **No clarity regarding rules**\nIn the external data thread, winner Selim Serfebekov had posed so many questions which begs the question of how clear the rules were in the end.\n+ **The licensing problems**\nIf you look at Giba's post in the forums - his team used CC-BY licensed data, which, correct me if I am wrong, is permitted here. So why did they have to use their second best submission and drop down the LB in the end?\n+ **Reproducibilty**\nI understand how necessary it is that these solutions are reproducible as they can go a long way - however with reproducibilty there was a blanket ban on using weights, that I had trained on Kaggle Kernels and uploaded to a dataset. Why\n\nI am thankful to Kaggle team, as I learnt a lot about image recognition and CV overall, but in a competition with the stakes so high, why did there have to be so much confusion?\n\nThe Kaggle team acts an intermediary after all - I feel that the hosts have the power regarding rules et cetera so I do not wish to proverbially shoot the messenger. So my question is to Mozaic and the other hosts - could you have made this a better experience for all of us? Would it be possible to?", "votes": null }, { "id": "885429", "postDate": "06/14/2020 07:56:04", "content": "I hope it is just because the weekend and kaggle will discuss the situation and share their point of view. Hopefully something more apologetic/constructive than
\n\n\n\n\nI am obliged to remind participants of the importance of reviewing the rules for each and every competition you wish to enter in its entirety, as the rules form a binding legal agreement between you and the competition sponsor. In particular, hosts call out their “Competition-Specific Rules” in Section A. That said, this experience reminds us that we can better advocate for improving participants’ understanding of these rules.
\n
I think with the Zillow issue @antgoldbloom gave quite a good response (https://www.kaggle.com/c/zillow-prize-1/discussion/45770). Right now I still have hope for compensation to the All Faces Are Real team.
\n\nLater we could discuss how to avoid such issues in the future and how to rebuild trust...
", "rawMarkdown": "I hope it is just because the weekend and kaggle will discuss the situation and share their point of view. Hopefully something more apologetic/constructive than \n\n> I am obliged to remind participants of the importance of reviewing the rules for each and every competition you wish to enter in its entirety, as the rules form a binding legal agreement between you and the competition sponsor. In particular, hosts call out their “Competition-Specific Rules” in Section A. That said, this experience reminds us that we can better advocate for improving participants’ understanding of these rules.\n\nI think with the Zillow issue @antgoldbloom gave quite a good response (https://www.kaggle.com/c/zillow-prize-1/discussion/45770). Right now I still have hope for compensation to the All Faces Are Real team.\n\nLater we could discuss how to avoid such issues in the future and how to rebuild trust...", "votes": null }, { "id": "885647", "postDate": "06/14/2020 11:12:43", "content": "Just to highlight the importance of the issue:\n
https://www.kaggle.com/gaborfodor/daily-top-forum-topics
", "rawMarkdown": "Just to highlight the importance of the issue:\n\n\nhttps://www.kaggle.com/gaborfodor/daily-top-forum-topics", "votes": null }, { "id": "886578", "postDate": "06/15/2020 06:22:55", "content": "The issue is even more dominant if we focus on 2020.\n* On the left: PetFinder.my Contest: 1st Place Winner Disqualified\n* On the right: 1st Place Removed Solution - All Faces Are Real Team
\n\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F18102%2F7af1c9a09d27875d54ee3ea24d197d18%2FScreenshot%202020-06-15%20at%208.18.59.png?generation=1592202071270636&alt=media\")
That's interesting @gaborfodor, thanks for the plots. Looks like Kaggle community is aware of the situation.
", "rawMarkdown": "That's interesting @gaborfodor, thanks for the plots. Looks like Kaggle community is aware of the situation.", "votes": null } ], "comments": [ { "id": 885429, "author_name": "gaborfodor", "author_url": "", "post_date": "06/14/2020 07:56:04", "content": "I hope it is just because the weekend and kaggle will discuss the situation and share their point of view. Hopefully something more apologetic/constructive than
\n\n\n\n\nI am obliged to remind participants of the importance of reviewing the rules for each and every competition you wish to enter in its entirety, as the rules form a binding legal agreement between you and the competition sponsor. In particular, hosts call out their “Competition-Specific Rules” in Section A. That said, this experience reminds us that we can better advocate for improving participants’ understanding of these rules.
\n
I think with the Zillow issue @antgoldbloom gave quite a good response (https://www.kaggle.com/c/zillow-prize-1/discussion/45770). Right now I still have hope for compensation to the All Faces Are Real team.
\n\nLater we could discuss how to avoid such issues in the future and how to rebuild trust...
", "votes": null, "replies": [ { "id": 885647, "author_name": "gaborfodor", "author_url": "", "post_date": "06/14/2020 11:12:43", "content": "Just to highlight the importance of the issue:\n
https://www.kaggle.com/gaborfodor/daily-top-forum-topics
", "votes": null, "replies": [] }, { "id": 886578, "author_name": "gaborfodor", "author_url": "", "post_date": "06/15/2020 06:22:55", "content": "The issue is even more dominant if we focus on 2020.\n* On the left: PetFinder.my Contest: 1st Place Winner Disqualified\n* On the right: 1st Place Removed Solution - All Faces Are Real Team
\n\n
That's interesting @gaborfodor, thanks for the plots. Looks like Kaggle community is aware of the situation.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "884808": "Given what happened with the team All Faces are Real, I feel like we have a lot of things unanswered in this competition:-\n\n+ **No clarity regarding rules**\nIn the external data thread, winner Selim Serfebekov had posed so many questions which begs the question of how clear the rules were in the end.\n+ **The licensing problems**\nIf you look at Giba's post in the forums - his team used CC-BY licensed data, which, correct me if I am wrong, is permitted here. So why did they have to use their second best submission and drop down the LB in the end?\n+ **Reproducibilty**\nI understand how necessary it is that these solutions are reproducible as they can go a long way - however with reproducibilty there was a blanket ban on using weights, that I had trained on Kaggle Kernels and uploaded to a dataset. Why\n\nI am thankful to Kaggle team, as I learnt a lot about image recognition and CV overall, but in a competition with the stakes so high, why did there have to be so much confusion?\n\nThe Kaggle team acts an intermediary after all - I feel that the hosts have the power regarding rules et cetera so I do not wish to proverbially shoot the messenger. So my question is to Mozaic and the other hosts - could you have made this a better experience for all of us? Would it be possible to?", "885429": "I hope it is just because the weekend and kaggle will discuss the situation and share their point of view. Hopefully something more apologetic/constructive than \n\n> I am obliged to remind participants of the importance of reviewing the rules for each and every competition you wish to enter in its entirety, as the rules form a binding legal agreement between you and the competition sponsor. In particular, hosts call out their “Competition-Specific Rules” in Section A. That said, this experience reminds us that we can better advocate for improving participants’ understanding of these rules.\n\nI think with the Zillow issue @antgoldbloom gave quite a good response (https://www.kaggle.com/c/zillow-prize-1/discussion/45770). Right now I still have hope for compensation to the All Faces Are Real team.\n\nLater we could discuss how to avoid such issues in the future and how to rebuild trust...", "885647": "Just to highlight the importance of the issue:\n\n\nhttps://www.kaggle.com/gaborfodor/daily-top-forum-topics", "886578": "The issue is even more dominant if we focus on 2020.\n* On the left: [PetFinder.my Contest: 1st Place Winner Disqualified](https://www.kaggle.com/c/petfinder-adoption-prediction/discussion/125436)\n* On the right: [1st Place Removed Solution - All Faces Are Real Team](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/157983)\n\n", "887031": "That's interesting @gaborfodor, thanks for the plots. Looks like Kaggle community is aware of the situation." }, "source": "meta" }