In my experience of some competitions, sharing high score public notebook is not recommended towards the end of competitions, l.
\nI cannot find this information in kaggle official rule, but in some case, they are warned by host or participants.
\nDoes anyone know about this rule?
UPDATE: modify title following ravi's comment.
\n\n", "messages": [ { "id": 2707330, "postDate": "2024-03-20T12:57:55.600Z", "content": "This usually is applicable in the last week of the competition, usually after the merger deadline
\n
This usually is applicable in the last week of the competition, usually after the merger deadline @clearwaterkzk
", "rawMarkdown": "This usually is applicable in the last week of the competition, usually after the merger deadline @clearwaterkzk ", "votes": 5, "replies": [ { "id": 2707383, "postDate": "2024-03-20T13:16:53.817Z", "content": "Thanks for sharing. I think, it may narrow the solution.
", "rawMarkdown": "Thanks for sharing. I think, it may narrow the solution.", "votes": 2 } ] }, { "id": 2707329, "postDate": "2024-03-20T12:56:54.373Z", "content": "In my experience of some competitions, sharing high score public notebook is not recommended towards the end of competitions, l.
\nI cannot find this information in kaggle official rule, but in some case, they are warned by host or participants.
\nDoes anyone know about this rule?
UPDATE: modify title following ravi's comment.
\n\n", "rawMarkdown": "In my experience of some competitions, sharing high score public notebook is not recommended **towards the end of competitions, l~~ike two weeks before ends~~.**\nI cannot find this information in kaggle official rule, but in some case, they are warned by host or participants.\nDoes anyone know about this rule?\n\n**UPDATE: modify title following ravi's comment.**\n>This usually is applicable in the last week of the competition, usually after the merger deadline\n", "votes": 3 }, { "id": 2735232, "postDate": "2024-04-04T15:56:34.593Z", "content": "This usually is applicable in the last week of the competition, usually after the merger deadline
\n
Updating public notebook gets high score now. What do we think about this..
\nhttps://www.kaggle.com/competitions/hms-harmful-brain-activity-classification/discussion/491070
did mention above it could happen. I saw your comments in the notebook a week ago about the issue that has now been fixed. Can understand wanting to correct code, maybe could have been done before public notebooks closed for last week?
\nThink there are issues with train labels and how notebooks have used votes, etc. so time will tell how these do on private LB.
Good luck @clearwaterkzk !
", "rawMarkdown": "did mention [above](https://www.kaggle.com/competitions/hms-harmful-brain-activity-classification/discussion/485352#2708819) it could happen. I saw your comments in the notebook a week ago about the issue that has now been fixed. Can understand wanting to correct code, maybe could have been done before public notebooks closed for last week? \nThink there are issues with train labels and how notebooks have used votes, etc. so time will tell how these do on private LB.\n\nGood luck @clearwaterkzk !", "votes": 1 } ] }, { "id": 2708819, "postDate": "2024-03-21T08:32:44.443Z", "content": "As mentioned by Ravi - in the last week (and after team merger is closed) on the Code page there will be a message about not publishing high scoring notebooks. And you cannot create a notebook and then make it Public, staying Private is OK.
\nBut, people can update their existing notebooks and run them, update models, and score them - it happens.
\nThis is also true for what is posted in Discussion late in competition - late sharing of full solutions or datasets can still be done.
So that comes down to following guidelines. It is good you post this topic now for those that may not be aware.
", "rawMarkdown": "As mentioned by Ravi - in the last week (and after team merger is closed) on the Code page there will be a message about not publishing high scoring notebooks. And you cannot create a notebook and then make it Public, staying Private is OK. \nBut, people can update their existing notebooks and run them, update models, and score them - it happens. \nThis is also true for what is posted in Discussion late in competition - late sharing of full solutions or datasets can still be done.\n\nSo that comes down to following guidelines. It is good you post this topic now for those that may not be aware. ", "votes": 1, "replies": [ { "id": 2708847, "postDate": "2024-03-21T08:44:24.707Z", "content": "Yes, I agree, it is possible to update an old notebook and the changes will be visible. It seems to me that there is a miscalculation of Kaggle in this and it will probably be necessary to focus on this problem in the next surveys conducted by Kaggle.
", "rawMarkdown": "Yes, I agree, it is possible to update an old notebook and the changes will be visible. It seems to me that there is a miscalculation of Kaggle in this and it will probably be necessary to focus on this problem in the next surveys conducted by Kaggle.", "votes": 1, "replies": [ { "id": 2709237, "postDate": "2024-03-21T13:51:03.007Z", "content": "Do you mean that, in extreme example, 1 hours before the end of competition, the highest notebook is updated and score is in medal zones, people who find the update and just copy-submit will win ?
", "rawMarkdown": "Do you mean that, in extreme example, 1 hours before the end of competition, the highest notebook is updated and score is in medal zones, people who find the update and just copy-submit will win ? ", "votes": 1 }, { "id": 2709330, "postDate": "2024-03-21T15:13:38.853Z", "content": "It can and has happened that existing notebooks are updated in the final week. Kaggle is aware of this.
\nBut realistically, most public notebooks will not win prizes or even gold. Sometimes silver more likely bronze.
\nHere though, and in many research competitions where public LB high scores come from ensembles and various weight changes, etc. to overfit the public LB, rarely do they do well in the private LB. Single models may be more of a chance but doubt any that would be in medal zone are public here.
The community guideline you speak of is real. People frown on sharing high scoring notebooks as the competition deadline approaches. Especially in the last week of the competition. You will avoid the anger of other competitors.
\nBut there is no rule, you can technically do so (the public/private settings are always there).
", "rawMarkdown": "The community guideline you speak of is real. People frown on sharing high scoring notebooks as the competition deadline approaches. Especially in the last week of the competition. You will avoid the anger of other competitors.\n\nBut there is no rule, you can technically do so (the public/private settings are always there). ", "votes": 1, "replies": [ { "id": 2707418, "postDate": "2024-03-20T13:38:50.690Z", "content": "Thanks for detail.
", "rawMarkdown": "Thanks for detail.", "votes": 1 }, { "id": 2707577, "postDate": "2024-03-20T15:21:35.933Z", "content": "In the last week of the competition, you cannot post a notebook in any way, since this is technically impossible due to the end of the deadline for creating groups and the end of the deadline for publishing notebooks (April 3, 2024), which means it does not violate any rules of the site.
", "rawMarkdown": "In the last week of the competition, you cannot post a notebook in any way, since this is technically impossible due to the end of the deadline for creating groups and the end of the deadline for publishing notebooks (April 3, 2024), which means it does not violate any rules of the site.", "votes": 1, "replies": [ { "id": 2708405, "postDate": "2024-03-21T02:13:29.430Z", "content": "Hi @konstantinboyko,
\nThe birthweight competition (https://www.kaggle.com/competitions/prediction-interval-competition-i-birth-weight/overview) ends tomorrow, has ~90 fairly advanced competitors, and 5 have publicly posted new code in the last 48 hours. See for yourself:
\n\nWhile not illegal, because it was shared via Kaggle in the competition specific forums, it is frowned upon. But it can be done. Understanding the community norms is helpful to avoid the practice.
\nCheers.
", "rawMarkdown": "Hi @konstantinboyko,\n\nThe birthweight competition (https://www.kaggle.com/competitions/prediction-interval-competition-i-birth-weight/overview) ends tomorrow, has ~90 fairly advanced competitors, and 5 have publicly posted new code in the last 48 hours. See for yourself:\n\nhttps://www.kaggle.com/competitions/prediction-interval-competition-i-birth-weight/code?competitionId=65445&sortBy=dateRun&excludeNonAccessedDatasources=true\n\nWhile not illegal, because it was shared via Kaggle in the competition specific forums, it is frowned upon. But it can be done. Understanding the community norms is helpful to avoid the practice.\n\nCheers.", "votes": 1 }, { "id": 2708484, "postDate": "2024-03-21T03:40:10.220Z", "content": "What will be published through the forums in the last week, and I completely agree with you, is a violation of the rules of the competition. In such cases, we are obliged to draw the attention of the site administration to fraud in this competition by notifying them.
\nI'd like to report cheating in a Competition
\nCheating is not taken lightly. We monitor compliance during competitions, and spend a considerable amount of time at the close of each competition to review suspect activity and to remove people who have cheated from the leaderboard. When we believe we have sufficient evidence, we take action through removal, or even an account ban
\nWe also monitor and investigate moderation reports (plagiarism, voting rings, etc.) throughout the week, and take action as appropriate, which includes removing medals as well as full-out banning accounts.
\nWe sometimes hear feedback that the anti-cheating actions we take don't get enough visibility. There are a number of reasons for this, but a key reason is that we want to put as much focus as possible on those things that make Kaggle great. Please know, though, that we work hard behind the scenes to keep Kaggle professional and fair to all.
\nIf you believe you have evidence that suggests a team violated competition rules, please report it to the Competitions Compliance Team for a thorough investigation.
", "rawMarkdown": "What will be published through the forums in the last week, and I completely agree with you, is a violation of the rules of the competition. In such cases, we are obliged to draw the attention of the site administration to fraud in this competition by notifying them.\n\nI'd like to report cheating in a Competition\n\nCheating is not taken lightly. We monitor compliance during competitions, and spend a considerable amount of time at the close of each competition to review suspect activity and to remove people who have cheated from the leaderboard. When we believe we have sufficient evidence, we take action through removal, or even an account ban\n\nWe also monitor and investigate moderation reports (plagiarism, voting rings, etc.) throughout the week, and take action as appropriate, which includes removing medals as well as full-out banning accounts.\n\nWe sometimes hear feedback that the anti-cheating actions we take don't get enough visibility. There are a number of reasons for this, but a key reason is that we want to put as much focus as possible on those things that make Kaggle great. Please know, though, that we work hard behind the scenes to keep Kaggle professional and fair to all.\n\nIf you believe you have evidence that suggests a team violated competition rules, please report it to the Competitions Compliance Team for a thorough investigation." }, { "id": 2708759, "postDate": "2024-03-21T07:58:32.940Z", "content": "The competition you mentioned is 'community' (i.e. the kind of competition anyone can open whenever he wants), thus the rules are less strict if exists at all. In official competitions, it's a different story. @toddgardiner
", "rawMarkdown": "The competition you mentioned is 'community' (i.e. the kind of competition anyone can open whenever he wants), thus the rules are less strict if exists at all. In official competitions, it's a different story. @toddgardiner ", "votes": 1 }, { "id": 2708815, "postDate": "2024-03-21T08:31:03.513Z", "content": "I have not and will not participate in official competitions, since communication and learning are more important to me than sports awards. And I thank you for posting your notebook in the Stanford Ribonanza RNA Folding competition, it taught me a lot.
", "rawMarkdown": "I have not and will not participate in official competitions, since communication and learning are more important to me than sports awards. And I thank you for posting your notebook in the Stanford Ribonanza RNA Folding competition, it taught me a lot.", "votes": 2 }, { "id": 2709077, "postDate": "2024-03-21T11:55:25.700Z", "content": "This (HMS) is an official competition, so you ARE participating in such :)
\nThank you for your kind words. Ribonanza was a great competition.
Thanks for the answer. I thought that HMS were community competitions.
", "rawMarkdown": "Thanks for the answer. I thought that HMS were community competitions." }, { "id": 2709261, "postDate": "2024-03-21T14:15:58.790Z", "content": "Not at all, look at the to line 'HARVARD MEDICAL SCHOOL · RESEARCH CODE COMPETITION · 18 DAYS TO GO'. For community comp' like the birthweight mentioned before it is 'CARL MCBRIDE ELLIS · COMMUNITY PREDICTION COMPETITION · 10 HOURS TO GO'
", "rawMarkdown": "Not at all, look at the to line 'HARVARD MEDICAL SCHOOL · RESEARCH CODE COMPETITION · 18 DAYS TO GO'. For community comp' like the birthweight mentioned before it is 'CARL MCBRIDE ELLIS · COMMUNITY PREDICTION COMPETITION · 10 HOURS TO GO'" } ] } ] }, { "id": 2709283, "postDate": "2024-03-21T14:34:06.723Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ], "comments": [ { "id": 2707330, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2024-03-20T12:57:55.600000", "content": "This usually is applicable in the last week of the competition, usually after the merger deadline @clearwaterkzk
", "votes": 5, "replies": [ { "id": 2707383, "author_name": "Aurora_blue", "author_url": "", "post_date": "2024-03-20T13:16:53.817000", "content": "Thanks for sharing. I think, it may narrow the solution.
", "votes": 2, "replies": [] } ] }, { "id": 2735232, "author_name": "Aurora_blue", "author_url": "", "post_date": "2024-04-04T15:56:34.593000", "content": "Updating public notebook gets high score now. What do we think about this..
\nhttps://www.kaggle.com/competitions/hms-harmful-brain-activity-classification/discussion/491070
did mention above it could happen. I saw your comments in the notebook a week ago about the issue that has now been fixed. Can understand wanting to correct code, maybe could have been done before public notebooks closed for last week?
\nThink there are issues with train labels and how notebooks have used votes, etc. so time will tell how these do on private LB.
Good luck @clearwaterkzk !
", "votes": 1, "replies": [] } ] }, { "id": 2708819, "author_name": "something4kag", "author_url": "", "post_date": "2024-03-21T08:32:44.443000", "content": "As mentioned by Ravi - in the last week (and after team merger is closed) on the Code page there will be a message about not publishing high scoring notebooks. And you cannot create a notebook and then make it Public, staying Private is OK.
\nBut, people can update their existing notebooks and run them, update models, and score them - it happens.
\nThis is also true for what is posted in Discussion late in competition - late sharing of full solutions or datasets can still be done.
So that comes down to following guidelines. It is good you post this topic now for those that may not be aware.
", "votes": 1, "replies": [ { "id": 2708847, "author_name": "AI-Cat", "author_url": "", "post_date": "2024-03-21T08:44:24.707000", "content": "Yes, I agree, it is possible to update an old notebook and the changes will be visible. It seems to me that there is a miscalculation of Kaggle in this and it will probably be necessary to focus on this problem in the next surveys conducted by Kaggle.
", "votes": 1, "replies": [ { "id": 2709237, "author_name": "Aurora_blue", "author_url": "", "post_date": "2024-03-21T13:51:03.007000", "content": "Do you mean that, in extreme example, 1 hours before the end of competition, the highest notebook is updated and score is in medal zones, people who find the update and just copy-submit will win ?
", "votes": 1, "replies": [] }, { "id": 2709330, "author_name": "something4kag", "author_url": "", "post_date": "2024-03-21T15:13:38.853000", "content": "It can and has happened that existing notebooks are updated in the final week. Kaggle is aware of this.
\nBut realistically, most public notebooks will not win prizes or even gold. Sometimes silver more likely bronze.
\nHere though, and in many research competitions where public LB high scores come from ensembles and various weight changes, etc. to overfit the public LB, rarely do they do well in the private LB. Single models may be more of a chance but doubt any that would be in medal zone are public here.
The community guideline you speak of is real. People frown on sharing high scoring notebooks as the competition deadline approaches. Especially in the last week of the competition. You will avoid the anger of other competitors.
\nBut there is no rule, you can technically do so (the public/private settings are always there).
", "votes": 1, "replies": [ { "id": 2707418, "author_name": "Aurora_blue", "author_url": "", "post_date": "2024-03-20T13:38:50.690000", "content": "Thanks for detail.
", "votes": 1, "replies": [] }, { "id": 2707577, "author_name": "AI-Cat", "author_url": "", "post_date": "2024-03-20T15:21:35.933000", "content": "In the last week of the competition, you cannot post a notebook in any way, since this is technically impossible due to the end of the deadline for creating groups and the end of the deadline for publishing notebooks (April 3, 2024), which means it does not violate any rules of the site.
", "votes": 1, "replies": [ { "id": 2708405, "author_name": "Todd Gardiner", "author_url": "", "post_date": "2024-03-21T02:13:29.430000", "content": "Hi @konstantinboyko,
\nThe birthweight competition (https://www.kaggle.com/competitions/prediction-interval-competition-i-birth-weight/overview) ends tomorrow, has ~90 fairly advanced competitors, and 5 have publicly posted new code in the last 48 hours. See for yourself:
\n\nWhile not illegal, because it was shared via Kaggle in the competition specific forums, it is frowned upon. But it can be done. Understanding the community norms is helpful to avoid the practice.
\nCheers.
", "votes": 1, "replies": [] }, { "id": 2708484, "author_name": "AI-Cat", "author_url": "", "post_date": "2024-03-21T03:40:10.220000", "content": "What will be published through the forums in the last week, and I completely agree with you, is a violation of the rules of the competition. In such cases, we are obliged to draw the attention of the site administration to fraud in this competition by notifying them.
\nI'd like to report cheating in a Competition
\nCheating is not taken lightly. We monitor compliance during competitions, and spend a considerable amount of time at the close of each competition to review suspect activity and to remove people who have cheated from the leaderboard. When we believe we have sufficient evidence, we take action through removal, or even an account ban
\nWe also monitor and investigate moderation reports (plagiarism, voting rings, etc.) throughout the week, and take action as appropriate, which includes removing medals as well as full-out banning accounts.
\nWe sometimes hear feedback that the anti-cheating actions we take don't get enough visibility. There are a number of reasons for this, but a key reason is that we want to put as much focus as possible on those things that make Kaggle great. Please know, though, that we work hard behind the scenes to keep Kaggle professional and fair to all.
\nIf you believe you have evidence that suggests a team violated competition rules, please report it to the Competitions Compliance Team for a thorough investigation.
", "votes": 0, "replies": [] }, { "id": 2708759, "author_name": "greySnow", "author_url": "", "post_date": "2024-03-21T07:58:32.940000", "content": "The competition you mentioned is 'community' (i.e. the kind of competition anyone can open whenever he wants), thus the rules are less strict if exists at all. In official competitions, it's a different story. @toddgardiner
", "votes": 1, "replies": [] }, { "id": 2708815, "author_name": "AI-Cat", "author_url": "", "post_date": "2024-03-21T08:31:03.513000", "content": "I have not and will not participate in official competitions, since communication and learning are more important to me than sports awards. And I thank you for posting your notebook in the Stanford Ribonanza RNA Folding competition, it taught me a lot.
", "votes": 2, "replies": [] }, { "id": 2709077, "author_name": "greySnow", "author_url": "", "post_date": "2024-03-21T11:55:25.700000", "content": "This (HMS) is an official competition, so you ARE participating in such :)
\nThank you for your kind words. Ribonanza was a great competition.
Thanks for the answer. I thought that HMS were community competitions.
", "votes": 0, "replies": [] }, { "id": 2709261, "author_name": "greySnow", "author_url": "", "post_date": "2024-03-21T14:15:58.790000", "content": "Not at all, look at the to line 'HARVARD MEDICAL SCHOOL · RESEARCH CODE COMPETITION · 18 DAYS TO GO'. For community comp' like the birthweight mentioned before it is 'CARL MCBRIDE ELLIS · COMMUNITY PREDICTION COMPETITION · 10 HOURS TO GO'
", "votes": 0, "replies": [] } ] } ] }, { "id": 2709283, "author_name": "", "author_url": "", "post_date": "2024-03-21T14:34:06.723000", "content": "", "votes": -1, "replies": [] } ], "raw_markdown_by_id": { "2707330": "This usually is applicable in the last week of the competition, usually after the merger deadline @clearwaterkzk ", "2707329": "In my experience of some competitions, sharing high score public notebook is not recommended **towards the end of competitions, l~~ike two weeks before ends~~.**\nI cannot find this information in kaggle official rule, but in some case, they are warned by host or participants.\nDoes anyone know about this rule?\n\n**UPDATE: modify title following ravi's comment.**\n>This usually is applicable in the last week of the competition, usually after the merger deadline\n", "2735232": "Updating public notebook gets high score now. What do we think about this..\nhttps://www.kaggle.com/competitions/hms-harmful-brain-activity-classification/discussion/491070", "2708819": "As mentioned by Ravi - in the last week (and after team merger is closed) on the Code page there will be a message about not publishing high scoring notebooks. And you cannot create a notebook and then make it Public, staying Private is OK. \nBut, people can update their existing notebooks and run them, update models, and score them - it happens. \nThis is also true for what is posted in Discussion late in competition - late sharing of full solutions or datasets can still be done.\n\nSo that comes down to following guidelines. It is good you post this topic now for those that may not be aware. ", "2707397": "The community guideline you speak of is real. People frown on sharing high scoring notebooks as the competition deadline approaches. Especially in the last week of the competition. You will avoid the anger of other competitors.\n\nBut there is no rule, you can technically do so (the public/private settings are always there). ", "2709283": "" } }