![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F769693%2F2a22b9652e0f7be981695afe2930a953%2F20201108020051.jpg?generation=1604772075034874&alt=media\")
Kaggler doesn't have to worry too much about this bug. It's a rare, very good competion. I have two suggestions for fixing this bug.
\nCheck every time env.predict () submitted row_ ID and iter_test provides exactly the same.
Prohibit modifying the generated submission.
I think it should be much better.
\nIf the organizer can delete my abnormal score that I can experience the competion normally, I would be very grateful.
", "messages": [ { "id": 1071973, "postDate": "2020-11-07T17:04:46.853Z", "content": "
Kaggler doesn't have to worry too much about this bug. It's a rare, very good competion. I have two suggestions for fixing this bug.
\nCheck every time env.predict () submitted row_ ID and iter_test provides exactly the same.
Prohibit modifying the generated submission.
I think it should be much better.
\nIf the organizer can delete my abnormal score that I can experience the competion normally, I would be very grateful.
", "rawMarkdown": "\n\n\nKaggler doesn't have to worry too much about this bug. It's a rare, very good competion. I have two suggestions for fixing this bug.\n\n1. Check every time env.predict () submitted row_ ID and iter_test provides exactly the same.\n\n2. Prohibit modifying the generated submission.\n\nI think it should be much better.\n\nIf the organizer can delete my abnormal score that I can experience the competion normally, I would be very grateful.", "votes": 223 }, { "id": 1072788, "postDate": "2020-11-08T17:26:33.173Z", "content": "Hi there,
\nThanks for raising this. We're thankful that the vast majority of the Kaggle community raises their concerns instead of simply try to exploit them for a win. We've received a number of emails or comments on previous competitions as well when a user believes they have uncovered something - it shows the integrity of our community and their desire to learn and grow instead of simply \"win.\"
\nThe Kaggle and Riiid teams will take a look at the bug and discuss options for what the competition can look like after a fix.
\nBest,
\nKaggle Team
", "rawMarkdown": "Hi there,\n\nThanks for raising this. We're thankful that the vast majority of the Kaggle community raises their concerns instead of simply try to exploit them for a win. We've received a number of emails or comments on previous competitions as well when a user believes they have uncovered something - it shows the integrity of our community and their desire to learn and grow instead of simply \"win.\"\n\nThe Kaggle and Riiid teams will take a look at the bug and discuss options for what the competition can look like after a fix.\n\nBest,\n\nKaggle Team", "votes": 30, "replies": [ { "id": 1073107, "postDate": "2020-11-09T06:07:52.607Z", "content": "Additional update posted here.
", "rawMarkdown": "Additional update [posted here.](https://www.kaggle.com/c/riiid-test-answer-prediction/discussion/196066)", "votes": 1 }, { "id": 1073124, "postDate": "2020-11-09T06:44:16.333Z", "content": "Thanks @sohier !
\n PS - The link is broken (missing a 'w' in 'www')
Fixed, thanks!
", "rawMarkdown": "Fixed, thanks!", "votes": 1 } ] }, { "id": 1072054, "postDate": "2020-11-07T18:43:27.940Z", "content": "Congrats, you are the winner and get $50,000. I left this competition.
", "rawMarkdown": "Congrats, you are the winner and get $50,000. I left this competition.", "votes": 12, "replies": [ { "id": 1072157, "postDate": "2020-11-07T20:56:16.733Z", "content": "A special award should be awarded to this author!
", "rawMarkdown": "A special award should be awarded to this author!", "votes": 14 }, { "id": 1087581, "postDate": "2020-11-22T21:48:33.230Z", "content": "Yes!
\nGood idea actually. Kaggle could have a badge like issues reporter or something …
\nOr the competition host can just create a special prize!
\nWe need to value more this good attitudes.
What are the guarantees that someone did not take advantage of this to get a more realistic score?
", "rawMarkdown": "What are the guarantees that someone did not take advantage of this to get a more realistic score?", "votes": 7, "replies": [ { "id": 1072077, "postDate": "2020-11-07T19:16:28.120Z", "content": "Good point, I think the solution is to deeply check all solutions manually at least for top teams at the end of this competition.
", "rawMarkdown": "Good point, I think the solution is to deeply check all solutions manually at least for top teams at the end of this competition.", "votes": 4 }, { "id": 1072079, "postDate": "2020-11-07T19:25:18.087Z", "content": "Given this, maybe a total change of all train and test data would be necessary.
", "rawMarkdown": "Given this, maybe a total change of all train and test data would be necessary." }, { "id": 1072083, "postDate": "2020-11-07T19:26:51.083Z", "content": "I don't think they are going to do that. Unfortunately the LB has to be restarted for this competition to make sense now…
", "rawMarkdown": "I don't think they are going to do that. Unfortunately the LB has to be restarted for this competition to make sense now...", "votes": 1 }, { "id": 1072084, "postDate": "2020-11-07T19:30:04.207Z", "content": "Would be possible to fix the API and re-run successful submitted notebooks?
", "rawMarkdown": "Would be possible to fix the API and re-run successful submitted notebooks? ", "votes": 5 }, { "id": 1072096, "postDate": "2020-11-07T19:48:29.190Z", "content": "It could also be possible that someone took advantage of this to explore the test set, maybe? Then it's hard to detect it even the scores would be rescored.
", "rawMarkdown": "It could also be possible that someone took advantage of this to explore the test set, maybe? Then it's hard to detect it even the scores would be rescored.", "votes": 1 }, { "id": 1072114, "postDate": "2020-11-07T20:17:02.207Z", "content": "Does this bug really help explore the test set? I can't think of a way to do it. I think kaggle team @sohier should fix API and re-run successful submitted notebook, then we'll have no problem.
", "rawMarkdown": "Does this bug really help explore the test set? I can't think of a way to do it. I think kaggle team @sohier should fix API and re-run successful submitted notebook, then we'll have no problem.", "votes": 16 }, { "id": 1072135, "postDate": "2020-11-07T20:35:06.977Z", "content": "Probably not. The kaggle team should investigate the cause of the bug and make sure unintended information didn't leak. I hope it's just a problem of the API.
", "rawMarkdown": "Probably not. The kaggle team should investigate the cause of the bug and make sure unintended information didn't leak. I hope it's just a problem of the API.", "votes": 1 }, { "id": 1072589, "postDate": "2020-11-08T13:27:29.503Z", "content": "Agree,,, great of @piupiupiu to share publically
\nHope the organisers change so that anyone else who benefitted and take advantage may not have the benefit… hoping these are removed as @batyazhizni suggest
@piupiupiu It's great that you were transparent with this exploit. If you found other bugs, hope you'll share them with the organizers so they can get them patched.
\n@sohier @hoonpyotimjeon @jineonjinbaek
", "rawMarkdown": "@piupiupiu It's great that you were transparent with this exploit. If you found other bugs, hope you'll share them with the organizers so they can get them patched. \n\n@sohier @hoonpyotimjeon @jineonjinbaek ", "votes": 8 }, { "id": 1071980, "postDate": "2020-11-07T17:19:52.217Z", "content": "Disappointing to see this. Submitting an empty dataframe as prediction is not a difficult edge case to come up with and should have been tested.
", "rawMarkdown": "Disappointing to see this. Submitting an empty dataframe as prediction is not a difficult edge case to come up with and should have been tested.", "votes": 5, "replies": [ { "id": 1071986, "postDate": "2020-11-07T17:25:15.313Z", "content": "I am not quite sure but it looks like the submitted dataframe isn't empty as that will happen for i=0 only, right? And i does get incremented at the very end by the chunk size.
\nNB code is bit out of indentation.
\n ........\n pre_test_df = test_df\n i += test_df.shape[0]\n ........\nSubmitting an empty dataframe is what allows you to get the perfect score. The API still thinks you're on the previous group even though you've moved on to the next one and have access to the labels for the previous group.
", "rawMarkdown": "Submitting an empty dataframe is what allows you to get the perfect score. The API still thinks you're on the previous group even though you've moved on to the next one and have access to the labels for the previous group.", "votes": 11 }, { "id": 1072008, "postDate": "2020-11-07T17:41:33.763Z", "content": "Ahh i understood it now, thanks Ming! No worries, this will be patched soon by the kaggle team.
", "rawMarkdown": "Ahh i understood it now, thanks Ming! No worries, this will be patched soon by the kaggle team." }, { "id": 1072028, "postDate": "2020-11-07T18:03:31.597Z", "content": "I hope this is fixed asap. It would be disappointing to see another exciting competition and hard work being destroyed by another bug :/
", "rawMarkdown": "I hope this is fixed asap. It would be disappointing to see another exciting competition and hard work being destroyed by another bug :/", "votes": 2 }, { "id": 1072249, "postDate": "2020-11-08T01:07:36.627Z", "content": "Thanks for explaining this
", "rawMarkdown": "Thanks for explaining this" } ] }, { "id": 1072430, "postDate": "2020-11-08T07:57:02.737Z", "content": "u1s1 得分着实吓人
", "rawMarkdown": "u1s1 得分着实吓人", "votes": 3 }, { "id": 1075287, "postDate": "2020-11-11T14:51:10.147Z", "content": "why people left comments like test test test? I don't get it ….
Me neither
", "rawMarkdown": "Me neither", "votes": 1 }, { "id": 1076900, "postDate": "2020-11-13T02:51:11.927Z", "content": "Yeah! me neither
", "rawMarkdown": "Yeah! me neither" }, { "id": 1077646, "postDate": "2020-11-13T19:05:50.530Z", "content": "They comment none-sense so as to boost as a \"Novice\" to a \"Contributor\". One of its requirements is to leave one comment.
", "rawMarkdown": "They comment none-sense so as to boost as a \"Novice\" to a \"Contributor\". One of its requirements is to leave one comment.", "votes": 2 }, { "id": 1079508, "postDate": "2020-11-16T06:21:42.747Z", "content": "yeah that is true
", "rawMarkdown": "yeah that is true" } ] }, { "id": 1074744, "postDate": "2020-11-11T02:32:03.093Z", "content": "Very Good Correctness Prediction
", "rawMarkdown": "Very Good Correctness Prediction", "votes": 4, "replies": [ { "id": 1160686, "postDate": "2021-01-20T04:30:38.913Z", "content": "Ok I See You Very Good Dissuction
", "rawMarkdown": "Ok I See You Very Good Dissuction", "votes": 3 } ] }, { "id": 1072069, "postDate": "2020-11-07T18:58:45.507Z", "content": "I really spent many hours here, I guess that's it !?
", "rawMarkdown": "I really spent many hours here, I guess that's it !? ", "votes": 4, "replies": [ { "id": 1072173, "postDate": "2020-11-07T21:23:59.223Z", "content": "If they patch the API to prevent this exploit then the competition is still on.
", "rawMarkdown": "If they patch the API to prevent this exploit then the competition is still on.", "votes": 2 } ] }, { "id": 1072308, "postDate": "2020-11-08T04:11:44.437Z", "content": "the current batch get the right answers of prior batch’s questions. need prohibit any chance of resubmit pred answer for previous questions.
", "rawMarkdown": "the current batch get the right answers of prior batch’s questions. need prohibit any chance of resubmit pred answer for previous questions. ", "votes": 1 }, { "id": 1072258, "postDate": "2020-11-08T01:51:32.923Z", "content": "The last iter_test has no next 'prior_group_answers_correct' , and it submits the real model's predict.Why the auc is still 1?
", "rawMarkdown": "The last iter_test has no next 'prior_group_answers_correct' , and it submits the real model's predict.Why the auc is still 1?", "votes": 1 }, { "id": 1072134, "postDate": "2020-11-07T20:34:35.157Z", "content": "How about another 70%, can u guarantee that you get highest score from unseen data by using characteristic of seen data even it is not statistically related to?
", "rawMarkdown": "How about another 70%, can u guarantee that you get highest score from unseen data by using characteristic of seen data even it is not statistically related to?", "votes": 1 }, { "id": 1073083, "postDate": "2020-11-09T05:28:05.150Z", "content": "I think one potential solution is to first fix the bug, then make the test dataset public together with the target, and generate new test dataset which follows chronologically after the old test dataset.
", "rawMarkdown": "I think one potential solution is to first fix the bug, then make the test dataset public together with the target, and generate new test dataset which follows chronologically after the old test dataset.", "votes": 2 }, { "id": 1072309, "postDate": "2020-11-08T04:18:27.540Z", "content": "@piupiupiu thank you for sharing this bug so openly. Your honesty is really commendable! Hope this bug gets resolved soon.
", "rawMarkdown": "@piupiupiu thank you for sharing this bug so openly. Your honesty is really commendable! Hope this bug gets resolved soon. ", "votes": 2 }, { "id": 1074199, "postDate": "2020-11-10T11:36:35.103Z", "content": "Same problem
", "rawMarkdown": "Same problem" }, { "id": 1074231, "postDate": "2020-11-10T12:13:41.410Z", "content": "test test test
", "rawMarkdown": "test test test", "votes": -3 }, { "id": 1073300, "postDate": "2020-11-09T12:18:19.370Z", "content": "test test test
", "rawMarkdown": "test test test", "votes": -5 }, { "id": 1073186, "postDate": "2020-11-09T08:41:15.897Z", "content": "test test test
", "rawMarkdown": "test test test", "votes": -5 }, { "id": 1075277, "postDate": "2020-11-11T14:41:07.407Z", "content": "test test test
", "rawMarkdown": "test test test", "votes": -7 }, { "id": 1073589, "postDate": "2020-11-09T16:57:19.843Z", "content": "test test test
", "rawMarkdown": "test test test", "votes": -5 }, { "id": 1161194, "postDate": "2021-01-20T12:04:13.363Z", "content": "Great work
", "rawMarkdown": "Great work" }, { "id": 1081399, "postDate": "2020-11-17T03:50:43.813Z", "content": "same problem
", "rawMarkdown": "same problem" }, { "id": 1081039, "postDate": "2020-11-16T18:39:42.407Z", "content": "That is correct!
", "rawMarkdown": "That is correct!" }, { "id": 1079507, "postDate": "2020-11-16T06:21:22.863Z", "content": "Thankyou for the integrity
", "rawMarkdown": "Thankyou for the integrity" }, { "id": 1079506, "postDate": "2020-11-16T06:21:18.753Z", "content": "yes that is a fact
", "rawMarkdown": "yes that is a fact" }, { "id": 1077828, "postDate": "2020-11-14T01:11:38.280Z", "content": "Ok, If I understood right, the problem is that I can send the answer of the previous batch utilizing the current batch's information(prior_group_answers_correct)?
Has this been fixed?
", "rawMarkdown": "Has this been fixed?\n" }, { "id": 1074987, "postDate": "2020-11-11T09:45:53.770Z", "content": "same here
", "rawMarkdown": "same here " }, { "id": 1074831, "postDate": "2020-11-11T05:51:23.083Z", "content": "Great Work Man! I tried to catch the bug, but I failed.
\nClaps !!!
yea same for me
", "rawMarkdown": "yea same for me" }, { "id": 1074104, "postDate": "2020-11-10T08:57:34.673Z", "content": "Greatest! 👍👍👍👍
", "rawMarkdown": "Greatest! 👍👍👍👍" }, { "id": 1074031, "postDate": "2020-11-10T06:55:14.240Z", "content": "You scared me too. But thanks for catching bugs.
", "rawMarkdown": "You scared me too. But thanks for catching bugs." }, { "id": 1073641, "postDate": "2020-11-09T18:26:56.970Z", "content": "you guy just scared me，hahah
", "rawMarkdown": "you guy just scared me，hahah" }, { "id": 1073279, "postDate": "2020-11-09T11:24:46.410Z", "content": "You didn't mention the Error. If you can drop that we might figure it out better
", "rawMarkdown": "You didn't mention the Error. If you can drop that we might figure it out better" }, { "id": 1072948, "postDate": "2020-11-08T23:31:11.507Z", "content": "I hope the organizers will correct this mistake (many thanks to @piupiupiu for sharing!) and we will be able to continue. Similar situations have happened before in other competitions.
\nVery interesting competition! I am just preparing new solutions.
Brillant work.
", "rawMarkdown": "Brillant work." }, { "id": 1072433, "postDate": "2020-11-08T07:59:27.800Z", "content": "Awesome hacking skills you got there @piupiupiu. Great that you shared the exploit!
", "rawMarkdown": "Awesome hacking skills you got there @piupiupiu. Great that you shared the exploit!" }, { "id": 1071982, "postDate": "2020-11-07T17:22:36.377Z", "content": "If i am understanding it, you are submitting the labels of the next chunk into the current chunk for almost 2.4M rows?
\nAnd for the remaining 1M rows, you predicted .66?
Thank you so much for sharing the issue openly, and we apologize for any confusion this has caused. I will discuss this with @sohier as soon as possible and take actions to make sure that this bug won't be affecting the fair evaluation of the competition.
", "rawMarkdown": "Thank you so much for sharing the issue openly, and we apologize for any confusion this has caused. I will discuss this with @sohier as soon as possible and take actions to make sure that this bug won't be affecting the fair evaluation of the competition.", "votes": 22, "isDeleted": true, "replies": [ { "id": 1072306, "postDate": "2020-11-08T04:06:50.293Z", "content": "Thanks @jineonjinbaek . I suggest starting a new thread to discuss any impact this might have to the current state of the competition. Hope this gets resolved soon
", "rawMarkdown": "Thanks @jineonjinbaek . I suggest starting a new thread to discuss any impact this might have to the current state of the competition. Hope this gets resolved soon" } ] }, { "id": 1077053, "postDate": "2020-11-13T07:39:55.220Z", "content": "Great! Thanks for pointing out👍
", "rawMarkdown": "Great! Thanks for pointing out👍" }, { "id": 1074740, "postDate": "2020-11-11T02:22:45.153Z", "content": "Thanks a lot!!!
", "rawMarkdown": "Thanks a lot!!!" }, { "id": 1073944, "postDate": "2020-11-10T03:58:03.127Z", "content": "Thank you for sharing this
", "rawMarkdown": "Thank you for sharing this" } ], "comments": [ { "id": 1072788, "author_name": "Addison Howard", "author_url": "", "post_date": "2020-11-08T17:26:33.173000", "content": "Hi there,
\nThanks for raising this. We're thankful that the vast majority of the Kaggle community raises their concerns instead of simply try to exploit them for a win. We've received a number of emails or comments on previous competitions as well when a user believes they have uncovered something - it shows the integrity of our community and their desire to learn and grow instead of simply \"win.\"
\nThe Kaggle and Riiid teams will take a look at the bug and discuss options for what the competition can look like after a fix.
\nBest,
\nKaggle Team
", "votes": 30, "replies": [ { "id": 1073107, "author_name": "Sohier Dane", "author_url": "", "post_date": "2020-11-09T06:07:52.607000", "content": "Additional update posted here.
", "votes": 1, "replies": [] }, { "id": 1073124, "author_name": "Abhimanyu Dikshit", "author_url": "", "post_date": "2020-11-09T06:44:16.333000", "content": "Thanks @sohier !
\n PS - The link is broken (missing a 'w' in 'www')
Fixed, thanks!
", "votes": 1, "replies": [] } ] }, { "id": 1072054, "author_name": "mamas", "author_url": "", "post_date": "2020-11-07T18:43:27.940000", "content": "Congrats, you are the winner and get $50,000. I left this competition.
", "votes": 12, "replies": [ { "id": 1072157, "author_name": "Yih-Dar SHIEH", "author_url": "", "post_date": "2020-11-07T20:56:16.733000", "content": "A special award should be awarded to this author!
", "votes": 14, "replies": [] }, { "id": 1087581, "author_name": "Koubouratou IDJATON", "author_url": "", "post_date": "2020-11-22T21:48:33.230000", "content": "Yes!
\nGood idea actually. Kaggle could have a badge like issues reporter or something …
\nOr the competition host can just create a special prize!
\nWe need to value more this good attitudes.
What are the guarantees that someone did not take advantage of this to get a more realistic score?
", "votes": 7, "replies": [ { "id": 1072077, "author_name": "Ala Eddine Ayadi", "author_url": "", "post_date": "2020-11-07T19:16:28.120000", "content": "Good point, I think the solution is to deeply check all solutions manually at least for top teams at the end of this competition.
", "votes": 4, "replies": [] }, { "id": 1072079, "author_name": "Icfstat", "author_url": "", "post_date": "2020-11-07T19:25:18.087000", "content": "Given this, maybe a total change of all train and test data would be necessary.
", "votes": 0, "replies": [] }, { "id": 1072083, "author_name": "Nikola Bacic", "author_url": "", "post_date": "2020-11-07T19:26:51.083000", "content": "I don't think they are going to do that. Unfortunately the LB has to be restarted for this competition to make sense now…
", "votes": 1, "replies": [] }, { "id": 1072084, "author_name": "Claudio Verdú Ruiz", "author_url": "", "post_date": "2020-11-07T19:30:04.207000", "content": "Would be possible to fix the API and re-run successful submitted notebooks?
", "votes": 5, "replies": [] }, { "id": 1072096, "author_name": "Yu Tsumura", "author_url": "", "post_date": "2020-11-07T19:48:29.190000", "content": "It could also be possible that someone took advantage of this to explore the test set, maybe? Then it's hard to detect it even the scores would be rescored.
", "votes": 1, "replies": [] }, { "id": 1072114, "author_name": "mamas", "author_url": "", "post_date": "2020-11-07T20:17:02.207000", "content": "Does this bug really help explore the test set? I can't think of a way to do it. I think kaggle team @sohier should fix API and re-run successful submitted notebook, then we'll have no problem.
", "votes": 16, "replies": [] }, { "id": 1072135, "author_name": "Yu Tsumura", "author_url": "", "post_date": "2020-11-07T20:35:06.977000", "content": "Probably not. The kaggle team should investigate the cause of the bug and make sure unintended information didn't leak. I hope it's just a problem of the API.
", "votes": 1, "replies": [] }, { "id": 1072589, "author_name": "Kamal Das", "author_url": "", "post_date": "2020-11-08T13:27:29.503000", "content": "Agree,,, great of @piupiupiu to share publically
\nHope the organisers change so that anyone else who benefitted and take advantage may not have the benefit… hoping these are removed as @batyazhizni suggest
@piupiupiu It's great that you were transparent with this exploit. If you found other bugs, hope you'll share them with the organizers so they can get them patched.
\n@sohier @hoonpyotimjeon @jineonjinbaek
", "votes": 8, "replies": [] }, { "id": 1071980, "author_name": "Ming Pan", "author_url": "", "post_date": "2020-11-07T17:19:52.217000", "content": "Disappointing to see this. Submitting an empty dataframe as prediction is not a difficult edge case to come up with and should have been tested.
", "votes": 5, "replies": [ { "id": 1071986, "author_name": "Aditya Soni", "author_url": "", "post_date": "2020-11-07T17:25:15.313000", "content": "I am not quite sure but it looks like the submitted dataframe isn't empty as that will happen for i=0 only, right? And i does get incremented at the very end by the chunk size.
\nNB code is bit out of indentation.
\n ........\n pre_test_df = test_df\n i += test_df.shape[0]\n ........\nSubmitting an empty dataframe is what allows you to get the perfect score. The API still thinks you're on the previous group even though you've moved on to the next one and have access to the labels for the previous group.
", "votes": 11, "replies": [] }, { "id": 1072008, "author_name": "Aditya Soni", "author_url": "", "post_date": "2020-11-07T17:41:33.763000", "content": "Ahh i understood it now, thanks Ming! No worries, this will be patched soon by the kaggle team.
", "votes": 0, "replies": [] }, { "id": 1072028, "author_name": "Icfstat", "author_url": "", "post_date": "2020-11-07T18:03:31.597000", "content": "I hope this is fixed asap. It would be disappointing to see another exciting competition and hard work being destroyed by another bug :/
", "votes": 2, "replies": [] }, { "id": 1072249, "author_name": "LGreig", "author_url": "", "post_date": "2020-11-08T01:07:36.627000", "content": "Thanks for explaining this
", "votes": 0, "replies": [] } ] }, { "id": 1072430, "author_name": "AliceNCsyuk", "author_url": "", "post_date": "2020-11-08T07:57:02.737000", "content": "u1s1 得分着实吓人
", "votes": 3, "replies": [] }, { "id": 1075287, "author_name": "Yih-Dar SHIEH", "author_url": "", "post_date": "2020-11-11T14:51:10.147000", "content": "why people left comments like test test test? I don't get it ….
Me neither
", "votes": 1, "replies": [] }, { "id": 1076900, "author_name": "Chandra Deep Rathan K", "author_url": "", "post_date": "2020-11-13T02:51:11.927000", "content": "Yeah! me neither
", "votes": 0, "replies": [] }, { "id": 1077646, "author_name": "Fatemeh Sangin", "author_url": "", "post_date": "2020-11-13T19:05:50.530000", "content": "They comment none-sense so as to boost as a \"Novice\" to a \"Contributor\". One of its requirements is to leave one comment.
", "votes": 2, "replies": [] }, { "id": 1079508, "author_name": "Aditya Singh", "author_url": "", "post_date": "2020-11-16T06:21:42.747000", "content": "yeah that is true
", "votes": 0, "replies": [] } ] }, { "id": 1074744, "author_name": "SURIYA CHAYATUMMAGOON", "author_url": "", "post_date": "2020-11-11T02:32:03.093000", "content": "Very Good Correctness Prediction
", "votes": 4, "replies": [ { "id": 1160686, "author_name": "Knowledge_SC", "author_url": "", "post_date": "2021-01-20T04:30:38.913000", "content": "Ok I See You Very Good Dissuction
", "votes": 3, "replies": [] } ] }, { "id": 1072069, "author_name": "Ala Eddine Ayadi", "author_url": "", "post_date": "2020-11-07T18:58:45.507000", "content": "I really spent many hours here, I guess that's it !?
", "votes": 4, "replies": [ { "id": 1072173, "author_name": "GhostSkipper", "author_url": "", "post_date": "2020-11-07T21:23:59.223000", "content": "If they patch the API to prevent this exploit then the competition is still on.
", "votes": 2, "replies": [] } ] }, { "id": 1072308, "author_name": "xjz", "author_url": "", "post_date": "2020-11-08T04:11:44.437000", "content": "the current batch get the right answers of prior batch’s questions. need prohibit any chance of resubmit pred answer for previous questions.
", "votes": 1, "replies": [] }, { "id": 1072258, "author_name": "sunShining", "author_url": "", "post_date": "2020-11-08T01:51:32.923000", "content": "The last iter_test has no next 'prior_group_answers_correct' , and it submits the real model's predict.Why the auc is still 1?
", "votes": 1, "replies": [] }, { "id": 1072134, "author_name": "Kemal Emre Çolak", "author_url": "", "post_date": "2020-11-07T20:34:35.157000", "content": "How about another 70%, can u guarantee that you get highest score from unseen data by using characteristic of seen data even it is not statistically related to?
", "votes": 1, "replies": [] }, { "id": 1073083, "author_name": "Frank Pan", "author_url": "", "post_date": "2020-11-09T05:28:05.150000", "content": "I think one potential solution is to first fix the bug, then make the test dataset public together with the target, and generate new test dataset which follows chronologically after the old test dataset.
", "votes": 2, "replies": [] }, { "id": 1072309, "author_name": "Naresh Jagadeesan", "author_url": "", "post_date": "2020-11-08T04:18:27.540000", "content": "@piupiupiu thank you for sharing this bug so openly. Your honesty is really commendable! Hope this bug gets resolved soon.
", "votes": 2, "replies": [] }, { "id": 1074199, "author_name": "Abhinav Prem", "author_url": "", "post_date": "2020-11-10T11:36:35.103000", "content": "Same problem
", "votes": 0, "replies": [] }, { "id": 1074231, "author_name": "Wally456", "author_url": "", "post_date": "2020-11-10T12:13:41.410000", "content": "test test test
", "votes": -3, "replies": [] }, { "id": 1073300, "author_name": "Minhaj Reza Hasan", "author_url": "", "post_date": "2020-11-09T12:18:19.370000", "content": "test test test
", "votes": -5, "replies": [] }, { "id": 1073186, "author_name": "tylerpi", "author_url": "", "post_date": "2020-11-09T08:41:15.897000", "content": "test test test
", "votes": -5, "replies": [] }, { "id": 1075277, "author_name": "Om Ashish Mishra", "author_url": "", "post_date": "2020-11-11T14:41:07.407000", "content": "test test test
", "votes": -7, "replies": [] }, { "id": 1073589, "author_name": "Vishveshwarrier27", "author_url": "", "post_date": "2020-11-09T16:57:19.843000", "content": "test test test
", "votes": -5, "replies": [] }, { "id": 1161194, "author_name": "Arda Gundogdu", "author_url": "", "post_date": "2021-01-20T12:04:13.363000", "content": "Great work
", "votes": 0, "replies": [] }, { "id": 1081399, "author_name": "EMGM28", "author_url": "", "post_date": "2020-11-17T03:50:43.813000", "content": "same problem
", "votes": 0, "replies": [] }, { "id": 1081039, "author_name": "Jesus Ponce", "author_url": "", "post_date": "2020-11-16T18:39:42.407000", "content": "That is correct!
", "votes": 0, "replies": [] }, { "id": 1079507, "author_name": "rpsantosa_kaggle", "author_url": "", "post_date": "2020-11-16T06:21:22.863000", "content": "Thankyou for the integrity
", "votes": 0, "replies": [] }, { "id": 1079506, "author_name": "Aditya Singh", "author_url": "", "post_date": "2020-11-16T06:21:18.753000", "content": "yes that is a fact
", "votes": 0, "replies": [] }, { "id": 1077828, "author_name": "jwc", "author_url": "", "post_date": "2020-11-14T01:11:38.280000", "content": "Ok, If I understood right, the problem is that I can send the answer of the previous batch utilizing the current batch's information(prior_group_answers_correct)?
Has this been fixed?
", "votes": 0, "replies": [] }, { "id": 1074987, "author_name": "Yunus S", "author_url": "", "post_date": "2020-11-11T09:45:53.770000", "content": "same here
", "votes": 0, "replies": [] }, { "id": 1074831, "author_name": "Arpit Bhushan Sharma", "author_url": "", "post_date": "2020-11-11T05:51:23.083000", "content": "Great Work Man! I tried to catch the bug, but I failed.
\nClaps !!!