{ "id": 190594, "title": "Help stop Plagiarism activity", "url": "/competitions/lyft-motion-prediction-autonomous-vehicles/discussion/190594", "author_name": "The Brown Iceman", "post_date": "2020-10-12T13:50:57.404000", "votes": 30, "comment_count": 37, "views": 0, "content": "

Dear community,

\n

The plagiarism problem is solved for now in this competition. I have decided to erase everything in post section on 14/10/2020 taking advice from Kaggle team over email. He has received account penalty. This issue is closed.

\n

Regards,
\nThe Brown Iceman

", "messages": [ { "id": 1047344, "postDate": "2020-10-12T13:50:57.403Z", "content": "

Dear community,

\n

The plagiarism problem is solved for now in this competition. I have decided to erase everything in post section on 14/10/2020 taking advice from Kaggle team over email. He has received account penalty. This issue is closed.

\n

Regards,
\nThe Brown Iceman

", "rawMarkdown": "Dear community,\n\nThe plagiarism problem is solved for now in this competition. I have decided to erase everything in post section on 14/10/2020 taking advice from Kaggle team over email. He has received account penalty. This issue is closed. \n\nRegards,\nThe Brown Iceman", "votes": 30 }, { "id": 1048623, "postDate": "2020-10-13T16:50:27.080Z", "content": "

I suspect that there is some work that can be done by the Kaggle Compliance and Moderation Team to help resolve these issues. In the meantime, however, I want to draw your attention to Kaggle's Community Guidelines. There are three rules in particular that I want to highlight that are very central to maintaining a healthy but competitive atmosphere:

\n\n

My recommendation is that everybody in this thread reviews their public notebooks and their public discussion posts through the lens of these new guidelines. If content that you posted might be in violation of one of these rules you should edit your post to prevent penalties from potentially being applied to your account.

", "rawMarkdown": "I suspect that there is some work that can be done by the Kaggle Compliance and Moderation Team to help resolve these issues. In the meantime, however, I want to draw your attention to [Kaggle's Community Guidelines](https://www.kaggle.com/general/60914). There are three rules in particular that I want to highlight that are very central to maintaining a healthy but competitive atmosphere:\n\n - `Discuss and respect ideas. Don’t make it personal. We’re all here to learn and share ideas. When you have critical feedback, focus on the ideas that others are sharing, not the person.`\n\n - `Harassment and threats are unacceptable. Low-level harassment is still harassment. Even minor or subtle put downs set a negative tone in our community that will alienate others.`\n\n - `Don’t try to manipulate the progression system. Progression manipulation includes any behaviors that intend to game Kaggle’s reputation system. They harm the fairness and quality of the community and are forbidden on our platform. Off-topic self promotion, upvote collusion, plagiarism, and cheating are some (but not all) examples of progression manipulation.`\n\nMy recommendation is that everybody in this thread reviews their public notebooks and their public discussion posts through the lens of these new guidelines. If content that you posted might be in violation of one of these rules you should edit your post to prevent penalties from potentially being applied to your account.", "votes": 17, "replies": [ { "id": 1048627, "postDate": "2020-10-13T16:59:25.153Z", "content": "

Thank you for the post : )

", "rawMarkdown": "Thank you for the post : )", "votes": 2 }, { "id": 1048648, "postDate": "2020-10-13T17:18:41.900Z", "content": "

The whole post is a violation of \"Harassment and threats are unacceptable. Low-level harassment is still harassment. Even minor or subtle put downs set a negative tone in our community that will alienate others.\" and https://www.gesetze-im-internet.de/englisch_stgb/englisch_stgb.html#p1812

", "rawMarkdown": "The whole post is a violation of \"Harassment and threats are unacceptable. Low-level harassment is still harassment. Even minor or subtle put downs set a negative tone in our community that will alienate others.\" and https://www.gesetze-im-internet.de/englisch_stgb/englisch_stgb.html#p1812", "votes": -8 }, { "id": 1048850, "postDate": "2020-10-13T21:27:16.790Z", "content": "

I'm not a fan of \"callout\" threads--if you see something wrong, report it, and let the admins deal with it. Dramatic exposures are just asking for trouble. But sending negative email to someone's employer over a dispute on Kaggle with a complete stranger is really unacceptable and also creepy. I recommend that you move on from this thread and try to contribute to Kaggle in a different way

", "rawMarkdown": "I'm not a fan of \"callout\" threads--if you see something wrong, report it, and let the admins deal with it. Dramatic exposures are just asking for trouble. But sending negative email to someone's employer over a dispute on Kaggle with a complete stranger is really unacceptable and also creepy. I recommend that you move on from this thread and try to contribute to Kaggle in a different way", "votes": 8 }, { "id": 1048981, "postDate": "2020-10-14T02:38:24.123Z", "content": "

\"sending negative email to someone's employer\" is a false statement.

", "rawMarkdown": "\"sending negative email to someone's employer\" is a false statement.", "votes": -8 } ] }, { "id": 1048099, "postDate": "2020-10-13T07:44:36.787Z", "content": "

What a great start to my day to hear that @thebigd8ta got all his notebooks removed. Justice served!

\n

He probably will just repeat what he did, but clearly he is just wasting his time by doing that.

", "rawMarkdown": "What a great start to my day to hear that @thebigd8ta got all his notebooks removed. Justice served!\n\nHe probably will just repeat what he did, but clearly he is just wasting his time by doing that.", "votes": 10 }, { "id": 1047662, "postDate": "2020-10-12T20:24:18.103Z", "content": "

Seems like someone took care of him ;)
\nAll notebooks but one from him are deleted. He only kept the rank (maybe this will be taken care off later).

\n

\"see

\n

edit: that was quick
\n\"update\"

\n

Thanks, for taking care, kaggle team.

", "rawMarkdown": "Seems like someone took care of him ;)\nAll notebooks but one from him are deleted. He only kept the rank (maybe this will be taken care off later).\n\n![see here](https://i.imgur.com/GB1vjTL.png)\n\nedit: that was quick\n![update](https://i.imgur.com/Cqd0AIF.png)\n\nThanks, for taking care, kaggle team.", "votes": 7, "replies": [ { "id": 1047665, "postDate": "2020-10-12T20:27:15.303Z", "content": "

Wow…. Certainly took long enough. I've been reporting every single one, great to see that Kaggle takes care of its honourable users.
\nNo more rank now either.

", "rawMarkdown": "Wow.... Certainly took long enough. I've been reporting every single one, great to see that Kaggle takes care of its honourable users.\nNo more rank now either. ", "votes": 3 }, { "id": 1048038, "postDate": "2020-10-13T06:30:27.857Z", "content": "

Not friendly post ..
\nWhat bothered the poster is that it scored higher than his work. It ranked higher with better score because it is an ENSEMBLE of two models; Always performs better than a single solution on unknown test; as the solution would be tailored to dataset used in the training. Please do not upvote! its current rank is thirty third.
\nit was made public to show how you can improve by combining two models .. with arbitrary weights 10% and 90% .. original models were public and the result was public with NEVER USE and NEVER UPVOTE.
\nIt never showed up in any discussion and never claimed as a solution. Clearly marked as copies (Allowed!!) of two top models and have shown the combination as better than originals. People who saw it perform better upvoted it for no clear reason! its current rank is thirty third:
\n**Combining Lyft Multimode Models .. Please Don't use or upvote Using a publicly shared kernel output for experimentation only with a weight of 10%
\nimport pandas as pd, numpy as np
\npd.options.display.max_columns=305

\n

Combining Lyft Multimode Models .. **Please Don't use or upvote **

\n

Using a publicly shared kernel output for experimentation only with a weight of 10%

\n

paths = [
\n \"../input/lyft-prediction-with-multi-mode-confidence/submission.csv\", # multimode model
\n \"../input/lyft-fsd/submission.csv\"
\n]

\n

Combining two models to check if the result improves;

\n

Weighted average:

\n

10% first model in paths array … model chosen here because it is a multimode model

\n

The rest for the second model .. submision from an earlier version v49

\n

weights = [0.1, 0.9]
\nconf_cols = np.array([\"conf_0\", \"conf_1\", \"conf_2\"])
\nxy_cols = [[],[],[]]
\nfor i in range(50):
\n for j in range(3):
\n xy_cols[j].append(f\"coord_x{j}{i}\")
\n xy_cols[j].append(f\"coord_y{j}{i}\")
\nxy_cols[0][:10]
\nCOLUMNS = [\"timestamp\", \"track_id\"] + list(conf_cols) + xy_cols[0] + xy_cols[1] + xy_cols[2]
\nSorting and Combining: All columns are provided from two mentioned models above
\ndef sort_df(df, sort_timestamp_track_id=True):

\n

conf_orders = np.argsort(-df[conf_cols].values,1)
\n XY = np.stack([df[xy_cols[0]].values,df[xy_cols[1]].values, df[xy_cols[2]].values], axis=1)
\n XY = XY[np.arange(len(XY))[:, None], conf_orders]
\n df2 = pd.DataFrame(columns = COLUMNS)
\n df2[\"timestamp\"] = df[\"timestamp\"].values
\n df2[\"track_id\"] = df[\"track_id\"].values
\n df2[xy_cols[0] + xy_cols[1] + xy_cols[2]] = XY.reshape(-1,300)
\n df2[conf_cols] = df[conf_cols].values[np.arange(len(df))[:, None], conf_orders]

\n

if sort_timestamp_track_id:
\n df2.sort_values([\"timestamp\", \"track_id\"], inplace=True)
\n df2.reset_index(inplace=True, drop=True)
\n return df2
\ndf = None
\nfor path,w in zip(paths,weights):
\n print(w, path)
\n temp = pd.read_csv(path)
\n temp = sort_df(temp)
\n temp[COLUMNS[5:]] *= w
\n if df is None:
\n df = temp
\n else:
\n df[COLUMNS[2:]] += temp[COLUMNS[2:]]
\ndf[conf_cols] /= df[conf_cols].sum(1).values[:, None]
\nsample = pd.read_csv(\"../input/lyft-motion-prediction-autonomous-vehicles/multi_mode_sample_submission.csv\")
\ndf = sample[[\"timestamp\", \"track_id\"]].merge(df, on=[\"timestamp\", \"track_id\"])
\nsample.shape, df.shape

\n

Building the submission output

\n

df.to_csv(\"submission.csv\", index=False, float_format='%.6f')**

", "rawMarkdown": "\nNot friendly post .. \n\nWhat bothered the poster is that it scored higher than his work. It ranked higher with better score because it is an ENSEMBLE of two models; Always performs better than a single solution on unknown test; as the solution would be tailored to dataset used in the training. Please do not upvote! its current rank is thirty third.\n\nit was made public to show how you can improve by combining two models .. with arbitrary weights 10% and 90% .. original models were public and the result was public with NEVER USE and NEVER UPVOTE. \n\nIt never showed up in any discussion and never claimed as a solution. Clearly marked as copies (Allowed!!) of two top models and have shown the combination as better than originals. ****People who saw it perform better upvoted it for no clear reason!**** its current rank is thirty third:\n\n**Combining Lyft Multimode Models .. Please Don't use or upvote Using a publicly shared kernel output for experimentation only with a weight of 10%\n\nimport pandas as pd, numpy as np\npd.options.display.max_columns=305\n#Combining Lyft Multimode Models .. **Please Don't use or upvote **\n#Using a publicly shared kernel output for experimentation only with a weight of 10%\n\npaths = [\n \"../input/lyft-prediction-with-multi-mode-confidence/submission.csv\", # multimode model\n \"../input/lyft-fsd/submission.csv\"\n]\n\n# Combining two models to check if the result improves; \n# Weighted average: \n# 10% first model in paths array ... model chosen here because it is a multimode model\n# The rest for the second model .. submision from an earlier version *v49*\n\nweights = [0.1, 0.9]\nconf_cols = np.array([\"conf_0\", \"conf_1\", \"conf_2\"])\nxy_cols = [[],[],[]]\nfor i in range(50):\n for j in range(3):\n xy_cols[j].append(f\"coord_x{j}{i}\")\n xy_cols[j].append(f\"coord_y{j}{i}\")\nxy_cols[0][:10]\nCOLUMNS = [\"timestamp\", \"track_id\"] + list(conf_cols) + xy_cols[0] + xy_cols[1] + xy_cols[2]\nSorting and Combining: All columns are provided from two mentioned models above\ndef sort_df(df, sort_timestamp_track_id=True):\n \n conf_orders = np.argsort(-df[conf_cols].values,1)\n XY = np.stack([df[xy_cols[0]].values,df[xy_cols[1]].values, df[xy_cols[2]].values], axis=1)\n XY = XY[np.arange(len(XY))[:, None], conf_orders]\n\n df2 = pd.DataFrame(columns = COLUMNS)\n df2[\"timestamp\"] = df[\"timestamp\"].values\n df2[\"track_id\"] = df[\"track_id\"].values\n df2[xy_cols[0] + xy_cols[1] + xy_cols[2]] = XY.reshape(-1,300)\n df2[conf_cols] = df[conf_cols].values[np.arange(len(df))[:, None], conf_orders]\n \n if sort_timestamp_track_id:\n df2.sort_values([\"timestamp\", \"track_id\"], inplace=True)\n df2.reset_index(inplace=True, drop=True)\n return df2\n\ndf = None\nfor path,w in zip(paths,weights):\n print(w, path)\n temp = pd.read_csv(path)\n temp = sort_df(temp)\n temp[COLUMNS[5:]] *= w\n if df is None:\n df = temp\n else:\n df[COLUMNS[2:]] += temp[COLUMNS[2:]]\ndf[conf_cols] /= df[conf_cols].sum(1).values[:, None]\n\nsample = pd.read_csv(\"../input/lyft-motion-prediction-autonomous-vehicles/multi_mode_sample_submission.csv\")\n\ndf = sample[[\"timestamp\", \"track_id\"]].merge(df, on=[\"timestamp\", \"track_id\"])\nsample.shape, df.shape\n# Building the submission output \ndf.to_csv(\"submission.csv\", index=False, float_format='%.6f')**", "votes": -8 } ] }, { "id": 1048611, "postDate": "2020-10-13T16:40:02.033Z", "content": "

I think admins should deal with this

", "rawMarkdown": "\nI think admins should deal with this", "votes": 3 }, { "id": 1050195, "postDate": "2020-10-15T06:56:56.043Z", "content": "

I cant visit his profile anymore, seems like it got deleted or he deleted it:
\nhttps://www.kaggle.com/thebigd8ta

", "rawMarkdown": "I cant visit his profile anymore, seems like it got deleted or he deleted it: \nhttps://www.kaggle.com/thebigd8ta", "votes": 4 }, { "id": 1049776, "postDate": "2020-10-14T19:07:56.997Z", "content": "

Just hit this thread and man does it feel good to see this guy finally get what he had coming for a long time. Thanks Kaggle team for stopping users like this and hoping for some others to follow soon!

", "rawMarkdown": "Just hit this thread and man does it feel good to see this guy finally get what he had coming for a long time. Thanks Kaggle team for stopping users like this and hoping for some others to follow soon!\n", "votes": 4 }, { "id": 1048404, "postDate": "2020-10-13T13:06:40.673Z", "content": "

Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "rawMarkdown": "Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.", "votes": 4 }, { "id": 1047371, "postDate": "2020-10-12T14:26:21.287Z", "content": "

I can see the two notebooks are indeed copied from other users' ones.
\nThis is to say that I acknowledge what you're reporting but I don't know if this is allowed or not (clearly it's not \"ethical\")

\n

Vlad will know more for sure though :)

", "rawMarkdown": "I can see the two notebooks are indeed copied from other users' ones. \nThis is to say that I acknowledge what you're reporting but I don't know if this is allowed or not (clearly it's not \"ethical\") \n\nVlad will know more for sure though :) ", "votes": 4, "replies": [ { "id": 1047378, "postDate": "2020-10-12T14:34:57.780Z", "content": "

Thanks for acknowledgement @lucabergamini . Copying is definitely allowed. What this person is doing is republishing other people work as soon as original author publishes a new model/result/notebook. This should totally not be allowed.

", "rawMarkdown": "Thanks for acknowledgement @lucabergamini . Copying is definitely allowed. What this person is doing is republishing other people work as soon as original author publishes a new model/result/notebook. This should totally not be allowed. ", "votes": 2 }, { "id": 1048043, "postDate": "2020-10-13T06:35:10.040Z", "content": "

Combining two models and clearly marking it with \"DO NOT USE\" .. never cared about the upvote! none of the aforementioned kernels ever published in code … if copying your output csv bothers you; then do not publish it. Certainly your attitude is not friendly! its current rank is thirty third.

", "rawMarkdown": "Combining two models and clearly marking it with \"DO NOT USE\" .. never cared about the upvote! none of the aforementioned kernels ever published in code ... if copying your output csv bothers you; then do not publish it. Certainly your attitude is not friendly! its current rank is thirty third.", "votes": -10 }, { "id": 1048622, "postDate": "2020-10-13T16:49:23.100Z", "rawMarkdown": "", "votes": -2, "isDeleted": true } ] }, { "id": 1050178, "postDate": "2020-10-15T06:26:02.640Z", "content": "

cool…Appreciated

", "rawMarkdown": "cool...Appreciated", "votes": 1 }, { "id": 1048662, "postDate": "2020-10-13T17:37:18.167Z", "content": "

I have pics of his threat from email notification. (Deleted the pic as issue is solved)
\nEdit: I have decided to empty the comment and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "rawMarkdown": " I have pics of his threat from email notification. (Deleted the pic as issue is solved)\n\nEdit: I have decided to empty the comment and what I wrote on 14/10/2020 taking advice from Kaggle team over email.", "votes": 1, "replies": [ { "id": 1048675, "postDate": "2020-10-13T17:46:03.670Z", "content": "

These are NOT threats .. corrections of your false claims

", "rawMarkdown": "These are NOT threats .. corrections of your false claims", "votes": -11 } ] }, { "id": 1047904, "postDate": "2020-10-13T03:44:59.773Z", "content": "

nice work dude

", "rawMarkdown": "nice work dude", "votes": 1, "replies": [ { "id": 1048139, "postDate": "2020-10-13T08:25:35.347Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ] }, { "id": 1048546, "postDate": "2020-10-13T15:37:26.623Z", "content": "

The \"Lyft ensemble II\" code had only 4 votes during its 52 versions! It was public from the start to demonstrate average weighting of two models. You can argue lost votes with the voters who missed upvoting your work.

", "rawMarkdown": "The \"Lyft ensemble II\" code had only 4 votes during its 52 versions! It was public from the start to demonstrate average weighting of two models. **You can argue lost votes with the voters who missed upvoting your work.** ", "votes": -14, "replies": [ { "id": 1048552, "postDate": "2020-10-13T15:42:57.740Z", "content": "

@thebigd8ta That is not the argument you should be making. You forked others' code, made miniscule changes, like changing dropout by 0.0001 to achieve a slightly better score, then got hundreds of forks and votes. Now you're sending messages to a Kaggler's employer who called you out? I'm quite shocked, to say the least.

", "rawMarkdown": "@thebigd8ta That is not the argument you should be making. You forked others' code, made miniscule changes, like changing dropout by 0.0001 to achieve a slightly better score, then got hundreds of forks and votes. Now you're sending messages to a Kaggler's employer who called you out? I'm quite shocked, to say the least.", "votes": 4 }, { "id": 1048554, "postDate": "2020-10-13T15:43:44.260Z", "content": "

Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "rawMarkdown": "Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.", "votes": 3 }, { "id": 1048584, "postDate": "2020-10-13T16:09:55.817Z", "content": "

People do the forks so they can see what models being used in an ensemble. Again, you can argue your case with the forkers; why would anyone upvote an ensemble? No dropout hyperparameter in the code; changes only in the models:

\n

paths = [
\n\"../input/lyft-prediction-with-multi-mode-confidence/submission.csv\", # multimode model
\n\"../input/lyft-fsd/submission.csv\"
\n]

", "rawMarkdown": "People do the forks so they can see what models being used in an ensemble. Again, you can argue your case with the forkers; why would anyone upvote an ensemble? No dropout hyperparameter in the code; changes only in the models:\n\npaths = [\n\"../input/lyft-prediction-with-multi-mode-confidence/submission.csv\", # multimode model\n\"../input/lyft-fsd/submission.csv\"\n]", "votes": -11 }, { "id": 1048586, "postDate": "2020-10-13T16:11:44.937Z", "content": "

Abusive language again and in violation of the penal code https://www.gesetze-im-internet.de/englisch_stgb/englisch_stgb.html#p1812 !

\n

you mention your ex-CEO in the middle of incredibly abusive post.

", "rawMarkdown": "Abusive language again and in violation of the penal code https://www.gesetze-im-internet.de/englisch_stgb/englisch_stgb.html#p1812 ! \n\nyou mention your ex-CEO in the middle of incredibly abusive post.", "votes": -7 }, { "id": 1048595, "postDate": "2020-10-13T16:20:01.117Z", "content": "

\"\"

", "rawMarkdown": "![](https://i.pinimg.com/originals/80/25/96/802596ecf53b50875b76651ce03d4dd6.jpg)", "votes": 11 }, { "id": 1048600, "postDate": "2020-10-13T16:24:16.663Z", "content": "

Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "rawMarkdown": "Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.", "votes": 2 }, { "id": 1048606, "postDate": "2020-10-13T16:33:38.270Z", "content": "

\"\"

", "rawMarkdown": "![](https://memegenerator.net/img/instances/66637237.jpg)", "votes": 9 }, { "id": 1048614, "postDate": "2020-10-13T16:42:46.587Z", "rawMarkdown": "", "votes": -3, "isDeleted": true } ] }, { "id": 1048033, "postDate": "2020-10-13T06:26:09.553Z", "content": "

All kernels mentioned were ensembled and were clearly marked not to use or upvote! Made public so as to acknowledge original work and useful combinations … AGAIN clearly marked as \"DO NOT USE\" and certainly \"Do Not Upvote\"!!

", "rawMarkdown": "All kernels mentioned were ensembled and were clearly marked not to use or upvote! Made public so as to acknowledge original work and useful combinations ... AGAIN clearly marked as \"DO NOT USE\" and certainly \"Do Not Upvote\"!! ", "votes": -11, "replies": [ { "id": 1048048, "postDate": "2020-10-13T06:37:58.370Z", "content": "

You did NOT change a single line in many of your copy kernels. So no, you did not mark them as such. It would have needed those line additions …

", "rawMarkdown": "You did NOT change a single line in many of your copy kernels. So no, you did not mark them as such. It would have needed those line additions ...", "votes": 3 }, { "id": 1048051, "postDate": "2020-10-13T06:42:01.377Z", "content": "

Only the CSVs were ever used… and combined ; These are not kernels or solutions and CLEARLY marked as ensembles and NOT to be used; its current rank is thirty third.

", "rawMarkdown": "Only the CSVs were ever used... and combined ; These are not kernels or solutions and CLEARLY marked as ensembles and NOT to be used; its current rank is thirty third.", "votes": -7 }, { "id": 1048235, "postDate": "2020-10-13T10:06:50.230Z", "content": "

Why would someone public something which helps getting a high rank without any effort and think a mark of NOT TO USE would keep medal hunters away?

\n

Do you believe it holds them off?

", "rawMarkdown": "Why would someone public something which helps getting a high rank without any effort and think a mark of **NOT TO USE** would keep medal hunters away?\n\nDo you believe it holds them off?", "votes": 2 }, { "id": 1048266, "postDate": "2020-10-13T10:39:55.013Z", "content": "

You can mesh the models and post.. most posted notebooks are over-trained; just for having a short shot at fame! you always can do better and that bothered the poster.

", "rawMarkdown": "You can mesh the models and post.. most posted notebooks are over-trained; just for having a short shot at fame! you always can do better and that bothered the poster.", "votes": -7 } ] }, { "id": 1050689, "postDate": "2020-10-15T16:21:30.020Z", "content": "

Thanks, Appericated!

", "rawMarkdown": "Thanks, Appericated!", "votes": 1 } ], "comments": [ { "id": 1048623, "author_name": "Paul Mooney", "author_url": "", "post_date": "2020-10-13T16:50:27.080000", "content": "

I suspect that there is some work that can be done by the Kaggle Compliance and Moderation Team to help resolve these issues. In the meantime, however, I want to draw your attention to Kaggle's Community Guidelines. There are three rules in particular that I want to highlight that are very central to maintaining a healthy but competitive atmosphere:

\n\n

My recommendation is that everybody in this thread reviews their public notebooks and their public discussion posts through the lens of these new guidelines. If content that you posted might be in violation of one of these rules you should edit your post to prevent penalties from potentially being applied to your account.

", "votes": 17, "replies": [ { "id": 1048627, "author_name": "The Brown Iceman", "author_url": "", "post_date": "2020-10-13T16:59:25.153000", "content": "

Thank you for the post : )

", "votes": 2, "replies": [] }, { "id": 1048648, "author_name": "", "author_url": "", "post_date": "2020-10-13T17:18:41.900000", "content": "

The whole post is a violation of \"Harassment and threats are unacceptable. Low-level harassment is still harassment. Even minor or subtle put downs set a negative tone in our community that will alienate others.\" and https://www.gesetze-im-internet.de/englisch_stgb/englisch_stgb.html#p1812

", "votes": -8, "replies": [] }, { "id": 1048850, "author_name": "CoreyJamesLevinson", "author_url": "", "post_date": "2020-10-13T21:27:16.790000", "content": "

I'm not a fan of \"callout\" threads--if you see something wrong, report it, and let the admins deal with it. Dramatic exposures are just asking for trouble. But sending negative email to someone's employer over a dispute on Kaggle with a complete stranger is really unacceptable and also creepy. I recommend that you move on from this thread and try to contribute to Kaggle in a different way

", "votes": 8, "replies": [] }, { "id": 1048981, "author_name": "", "author_url": "", "post_date": "2020-10-14T02:38:24.123000", "content": "

\"sending negative email to someone's employer\" is a false statement.

", "votes": -8, "replies": [] } ] }, { "id": 1048099, "author_name": "Gilles Vandewiele", "author_url": "", "post_date": "2020-10-13T07:44:36.787000", "content": "

What a great start to my day to hear that @thebigd8ta got all his notebooks removed. Justice served!

\n

He probably will just repeat what he did, but clearly he is just wasting his time by doing that.

", "votes": 10, "replies": [] }, { "id": 1047662, "author_name": "Pascal Pfeiffer", "author_url": "", "post_date": "2020-10-12T20:24:18.103000", "content": "

Seems like someone took care of him ;)
\nAll notebooks but one from him are deleted. He only kept the rank (maybe this will be taken care off later).

\n

\"see

\n

edit: that was quick
\n\"update\"

\n

Thanks, for taking care, kaggle team.

", "votes": 7, "replies": [ { "id": 1047665, "author_name": "Stanley Zheng", "author_url": "", "post_date": "2020-10-12T20:27:15.303000", "content": "

Wow…. Certainly took long enough. I've been reporting every single one, great to see that Kaggle takes care of its honourable users.
\nNo more rank now either.

", "votes": 3, "replies": [] }, { "id": 1048038, "author_name": "", "author_url": "", "post_date": "2020-10-13T06:30:27.857000", "content": "

Not friendly post ..
\nWhat bothered the poster is that it scored higher than his work. It ranked higher with better score because it is an ENSEMBLE of two models; Always performs better than a single solution on unknown test; as the solution would be tailored to dataset used in the training. Please do not upvote! its current rank is thirty third.
\nit was made public to show how you can improve by combining two models .. with arbitrary weights 10% and 90% .. original models were public and the result was public with NEVER USE and NEVER UPVOTE.
\nIt never showed up in any discussion and never claimed as a solution. Clearly marked as copies (Allowed!!) of two top models and have shown the combination as better than originals. People who saw it perform better upvoted it for no clear reason! its current rank is thirty third:
\n**Combining Lyft Multimode Models .. Please Don't use or upvote Using a publicly shared kernel output for experimentation only with a weight of 10%
\nimport pandas as pd, numpy as np
\npd.options.display.max_columns=305

\n

Combining Lyft Multimode Models .. **Please Don't use or upvote **

\n

Using a publicly shared kernel output for experimentation only with a weight of 10%

\n

paths = [
\n \"../input/lyft-prediction-with-multi-mode-confidence/submission.csv\", # multimode model
\n \"../input/lyft-fsd/submission.csv\"
\n]

\n

Combining two models to check if the result improves;

\n

Weighted average:

\n

10% first model in paths array … model chosen here because it is a multimode model

\n

The rest for the second model .. submision from an earlier version v49

\n

weights = [0.1, 0.9]
\nconf_cols = np.array([\"conf_0\", \"conf_1\", \"conf_2\"])
\nxy_cols = [[],[],[]]
\nfor i in range(50):
\n for j in range(3):
\n xy_cols[j].append(f\"coord_x{j}{i}\")
\n xy_cols[j].append(f\"coord_y{j}{i}\")
\nxy_cols[0][:10]
\nCOLUMNS = [\"timestamp\", \"track_id\"] + list(conf_cols) + xy_cols[0] + xy_cols[1] + xy_cols[2]
\nSorting and Combining: All columns are provided from two mentioned models above
\ndef sort_df(df, sort_timestamp_track_id=True):

\n

conf_orders = np.argsort(-df[conf_cols].values,1)
\n XY = np.stack([df[xy_cols[0]].values,df[xy_cols[1]].values, df[xy_cols[2]].values], axis=1)
\n XY = XY[np.arange(len(XY))[:, None], conf_orders]
\n df2 = pd.DataFrame(columns = COLUMNS)
\n df2[\"timestamp\"] = df[\"timestamp\"].values
\n df2[\"track_id\"] = df[\"track_id\"].values
\n df2[xy_cols[0] + xy_cols[1] + xy_cols[2]] = XY.reshape(-1,300)
\n df2[conf_cols] = df[conf_cols].values[np.arange(len(df))[:, None], conf_orders]

\n

if sort_timestamp_track_id:
\n df2.sort_values([\"timestamp\", \"track_id\"], inplace=True)
\n df2.reset_index(inplace=True, drop=True)
\n return df2
\ndf = None
\nfor path,w in zip(paths,weights):
\n print(w, path)
\n temp = pd.read_csv(path)
\n temp = sort_df(temp)
\n temp[COLUMNS[5:]] *= w
\n if df is None:
\n df = temp
\n else:
\n df[COLUMNS[2:]] += temp[COLUMNS[2:]]
\ndf[conf_cols] /= df[conf_cols].sum(1).values[:, None]
\nsample = pd.read_csv(\"../input/lyft-motion-prediction-autonomous-vehicles/multi_mode_sample_submission.csv\")
\ndf = sample[[\"timestamp\", \"track_id\"]].merge(df, on=[\"timestamp\", \"track_id\"])
\nsample.shape, df.shape

\n

Building the submission output

\n

df.to_csv(\"submission.csv\", index=False, float_format='%.6f')**

", "votes": -8, "replies": [] } ] }, { "id": 1048611, "author_name": "Akobirkhuja", "author_url": "", "post_date": "2020-10-13T16:40:02.033000", "content": "

I think admins should deal with this

", "votes": 3, "replies": [] }, { "id": 1050195, "author_name": "Ali Abdin", "author_url": "", "post_date": "2020-10-15T06:56:56.043000", "content": "

I cant visit his profile anymore, seems like it got deleted or he deleted it:
\nhttps://www.kaggle.com/thebigd8ta

", "votes": 4, "replies": [] }, { "id": 1049776, "author_name": "Alex Bader", "author_url": "", "post_date": "2020-10-14T19:07:56.997000", "content": "

Just hit this thread and man does it feel good to see this guy finally get what he had coming for a long time. Thanks Kaggle team for stopping users like this and hoping for some others to follow soon!

", "votes": 4, "replies": [] }, { "id": 1048404, "author_name": "The Brown Iceman", "author_url": "", "post_date": "2020-10-13T13:06:40.673000", "content": "

Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "votes": 4, "replies": [] }, { "id": 1047371, "author_name": "Luca Bergamini", "author_url": "", "post_date": "2020-10-12T14:26:21.287000", "content": "

I can see the two notebooks are indeed copied from other users' ones.
\nThis is to say that I acknowledge what you're reporting but I don't know if this is allowed or not (clearly it's not \"ethical\")

\n

Vlad will know more for sure though :)

", "votes": 4, "replies": [ { "id": 1047378, "author_name": "The Brown Iceman", "author_url": "", "post_date": "2020-10-12T14:34:57.780000", "content": "

Thanks for acknowledgement @lucabergamini . Copying is definitely allowed. What this person is doing is republishing other people work as soon as original author publishes a new model/result/notebook. This should totally not be allowed.

", "votes": 2, "replies": [] }, { "id": 1048043, "author_name": "", "author_url": "", "post_date": "2020-10-13T06:35:10.040000", "content": "

Combining two models and clearly marking it with \"DO NOT USE\" .. never cared about the upvote! none of the aforementioned kernels ever published in code … if copying your output csv bothers you; then do not publish it. Certainly your attitude is not friendly! its current rank is thirty third.

", "votes": -10, "replies": [] }, { "id": 1048622, "author_name": "", "author_url": "", "post_date": "2020-10-13T16:49:23.100000", "content": "", "votes": -2, "replies": [] } ] }, { "id": 1050178, "author_name": "NamanPandey", "author_url": "", "post_date": "2020-10-15T06:26:02.640000", "content": "

cool…Appreciated

", "votes": 1, "replies": [] }, { "id": 1048662, "author_name": "The Brown Iceman", "author_url": "", "post_date": "2020-10-13T17:37:18.167000", "content": "

I have pics of his threat from email notification. (Deleted the pic as issue is solved)
\nEdit: I have decided to empty the comment and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "votes": 1, "replies": [ { "id": 1048675, "author_name": "", "author_url": "", "post_date": "2020-10-13T17:46:03.670000", "content": "

These are NOT threats .. corrections of your false claims

", "votes": -11, "replies": [] } ] }, { "id": 1047904, "author_name": "Mukul Singh", "author_url": "", "post_date": "2020-10-13T03:44:59.773000", "content": "

nice work dude

", "votes": 1, "replies": [ { "id": 1048139, "author_name": "", "author_url": "", "post_date": "2020-10-13T08:25:35.347000", "content": "", "votes": -1, "replies": [] } ] }, { "id": 1048546, "author_name": "", "author_url": "", "post_date": "2020-10-13T15:37:26.623000", "content": "

The \"Lyft ensemble II\" code had only 4 votes during its 52 versions! It was public from the start to demonstrate average weighting of two models. You can argue lost votes with the voters who missed upvoting your work.

", "votes": -14, "replies": [ { "id": 1048552, "author_name": "Stanley Zheng", "author_url": "", "post_date": "2020-10-13T15:42:57.740000", "content": "

@thebigd8ta That is not the argument you should be making. You forked others' code, made miniscule changes, like changing dropout by 0.0001 to achieve a slightly better score, then got hundreds of forks and votes. Now you're sending messages to a Kaggler's employer who called you out? I'm quite shocked, to say the least.

", "votes": 4, "replies": [] }, { "id": 1048554, "author_name": "The Brown Iceman", "author_url": "", "post_date": "2020-10-13T15:43:44.260000", "content": "

Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "votes": 3, "replies": [] }, { "id": 1048584, "author_name": "", "author_url": "", "post_date": "2020-10-13T16:09:55.817000", "content": "

People do the forks so they can see what models being used in an ensemble. Again, you can argue your case with the forkers; why would anyone upvote an ensemble? No dropout hyperparameter in the code; changes only in the models:

\n

paths = [
\n\"../input/lyft-prediction-with-multi-mode-confidence/submission.csv\", # multimode model
\n\"../input/lyft-fsd/submission.csv\"
\n]

", "votes": -11, "replies": [] }, { "id": 1048586, "author_name": "", "author_url": "", "post_date": "2020-10-13T16:11:44.937000", "content": "

Abusive language again and in violation of the penal code https://www.gesetze-im-internet.de/englisch_stgb/englisch_stgb.html#p1812 !

\n

you mention your ex-CEO in the middle of incredibly abusive post.

", "votes": -7, "replies": [] }, { "id": 1048595, "author_name": "Tucker Arrants", "author_url": "", "post_date": "2020-10-13T16:20:01.117000", "content": "

\"\"

", "votes": 11, "replies": [] }, { "id": 1048600, "author_name": "The Brown Iceman", "author_url": "", "post_date": "2020-10-13T16:24:16.663000", "content": "

Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.

", "votes": 2, "replies": [] }, { "id": 1048606, "author_name": "Pascal Pfeiffer", "author_url": "", "post_date": "2020-10-13T16:33:38.270000", "content": "

\"\"

", "votes": 9, "replies": [] }, { "id": 1048614, "author_name": "", "author_url": "", "post_date": "2020-10-13T16:42:46.587000", "content": "", "votes": -3, "replies": [] } ] }, { "id": 1048033, "author_name": "", "author_url": "", "post_date": "2020-10-13T06:26:09.553000", "content": "

All kernels mentioned were ensembled and were clearly marked not to use or upvote! Made public so as to acknowledge original work and useful combinations … AGAIN clearly marked as \"DO NOT USE\" and certainly \"Do Not Upvote\"!!

", "votes": -11, "replies": [ { "id": 1048048, "author_name": "Pascal Pfeiffer", "author_url": "", "post_date": "2020-10-13T06:37:58.370000", "content": "

You did NOT change a single line in many of your copy kernels. So no, you did not mark them as such. It would have needed those line additions …

", "votes": 3, "replies": [] }, { "id": 1048051, "author_name": "", "author_url": "", "post_date": "2020-10-13T06:42:01.377000", "content": "

Only the CSVs were ever used… and combined ; These are not kernels or solutions and CLEARLY marked as ensembles and NOT to be used; its current rank is thirty third.

", "votes": -7, "replies": [] }, { "id": 1048235, "author_name": "Ali Abdin", "author_url": "", "post_date": "2020-10-13T10:06:50.230000", "content": "

Why would someone public something which helps getting a high rank without any effort and think a mark of NOT TO USE would keep medal hunters away?

\n

Do you believe it holds them off?

", "votes": 2, "replies": [] }, { "id": 1048266, "author_name": "", "author_url": "", "post_date": "2020-10-13T10:39:55.013000", "content": "

You can mesh the models and post.. most posted notebooks are over-trained; just for having a short shot at fame! you always can do better and that bothered the poster.

", "votes": -7, "replies": [] } ] }, { "id": 1050689, "author_name": "Gaurav Chauhan", "author_url": "", "post_date": "2020-10-15T16:21:30.020000", "content": "

Thanks, Appericated!

", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "1047344": "Dear community,\n\nThe plagiarism problem is solved for now in this competition. I have decided to erase everything in post section on 14/10/2020 taking advice from Kaggle team over email. He has received account penalty. This issue is closed. \n\nRegards,\nThe Brown Iceman", "1048623": "I suspect that there is some work that can be done by the Kaggle Compliance and Moderation Team to help resolve these issues. In the meantime, however, I want to draw your attention to [Kaggle's Community Guidelines](https://www.kaggle.com/general/60914). There are three rules in particular that I want to highlight that are very central to maintaining a healthy but competitive atmosphere:\n\n - `Discuss and respect ideas. Don’t make it personal. We’re all here to learn and share ideas. When you have critical feedback, focus on the ideas that others are sharing, not the person.`\n\n - `Harassment and threats are unacceptable. Low-level harassment is still harassment. Even minor or subtle put downs set a negative tone in our community that will alienate others.`\n\n - `Don’t try to manipulate the progression system. Progression manipulation includes any behaviors that intend to game Kaggle’s reputation system. They harm the fairness and quality of the community and are forbidden on our platform. Off-topic self promotion, upvote collusion, plagiarism, and cheating are some (but not all) examples of progression manipulation.`\n\nMy recommendation is that everybody in this thread reviews their public notebooks and their public discussion posts through the lens of these new guidelines. If content that you posted might be in violation of one of these rules you should edit your post to prevent penalties from potentially being applied to your account.", "1048099": "What a great start to my day to hear that @thebigd8ta got all his notebooks removed. Justice served!\n\nHe probably will just repeat what he did, but clearly he is just wasting his time by doing that.", "1047662": "Seems like someone took care of him ;)\nAll notebooks but one from him are deleted. He only kept the rank (maybe this will be taken care off later).\n\n![see here](https://i.imgur.com/GB1vjTL.png)\n\nedit: that was quick\n![update](https://i.imgur.com/Cqd0AIF.png)\n\nThanks, for taking care, kaggle team.", "1048611": "\nI think admins should deal with this", "1050195": "I cant visit his profile anymore, seems like it got deleted or he deleted it: \nhttps://www.kaggle.com/thebigd8ta", "1049776": "Just hit this thread and man does it feel good to see this guy finally get what he had coming for a long time. Thanks Kaggle team for stopping users like this and hoping for some others to follow soon!\n", "1048404": "Edit: I have decided to empty the comment i had written and what I wrote on 14/10/2020 taking advice from Kaggle team over email.", "1047371": "I can see the two notebooks are indeed copied from other users' ones. \nThis is to say that I acknowledge what you're reporting but I don't know if this is allowed or not (clearly it's not \"ethical\") \n\nVlad will know more for sure though :) ", "1050178": "cool...Appreciated", "1048662": " I have pics of his threat from email notification. (Deleted the pic as issue is solved)\n\nEdit: I have decided to empty the comment and what I wrote on 14/10/2020 taking advice from Kaggle team over email.", "1047904": "nice work dude", "1048546": "The \"Lyft ensemble II\" code had only 4 votes during its 52 versions! It was public from the start to demonstrate average weighting of two models. **You can argue lost votes with the voters who missed upvoting your work.** ", "1048033": "All kernels mentioned were ensembled and were clearly marked not to use or upvote! Made public so as to acknowledge original work and useful combinations ... AGAIN clearly marked as \"DO NOT USE\" and certainly \"Do Not Upvote\"!! ", "1050689": "Thanks, Appericated!" } }