![\"image1\"](\"https://storage.googleapis.com/kaggle-forum-message-attachments/1118070/17721/copied2.JPG\")
A few days ago @yasufuminakama published two great notebooks for respectively training and submit a ResNext model, which at the time achieved the highest score on the leaderboard out of the public notebooks. They are:
\n\nIt has been a few days now and the notebooks are sitting at ~40 upvotes each, whereas usually top notebooks in medical imaging competitions would have gotten gold by now.
\nNow, this is understandable because there are many on-going competitions and people might not have time to compete during holidays. In fact, I wouldn't have been bothered if it wasn't for multiple public notebooks blatantly copying @yasufuminakama's inference notebook and changing the name so people wouldn't so easily recognize it.
\nIn fact, one notebook that made me very angry had:
\nI wouldn't be bothered by this if it was an honest mistake (e.g. accidentally made public after forking), but it is clear that the name change and erasing reference to the original notebook were not accidental or a beginner's mistake. What disappointed me is that many of the people that upvoted the plagiarized notebook didn't even upvote the original notebook, which indicates that even with the two indicators (ref. to the screenshots), it's still hard for people to find the original notebook and determine if there's any worthy change.
\n
![\"image2\"](\"https://storage.googleapis.com/kaggle-forum-message-attachments/1118070/17722/copied1.JPG\")
I don't want to start a witch hunt, instead I'd rather want everyone to go take a look at @yasufuminakama and give him an upvote if you appreciated them. Many kagglers are working hard on their notebooks and are very proud, and this unpunished plagiarism that gets rewarded with medals will only deter them to produce original work in the future.
", "messages": [ { "id": 1118070, "postDate": "2020-12-18T17:40:00.123Z", "content": "A few days ago @yasufuminakama published two great notebooks for respectively training and submit a ResNext model, which at the time achieved the highest score on the leaderboard out of the public notebooks. They are:
\n\nIt has been a few days now and the notebooks are sitting at ~40 upvotes each, whereas usually top notebooks in medical imaging competitions would have gotten gold by now.
\nNow, this is understandable because there are many on-going competitions and people might not have time to compete during holidays. In fact, I wouldn't have been bothered if it wasn't for multiple public notebooks blatantly copying @yasufuminakama's inference notebook and changing the name so people wouldn't so easily recognize it.
\nIn fact, one notebook that made me very angry had:
\nI wouldn't be bothered by this if it was an honest mistake (e.g. accidentally made public after forking), but it is clear that the name change and erasing reference to the original notebook were not accidental or a beginner's mistake. What disappointed me is that many of the people that upvoted the plagiarized notebook didn't even upvote the original notebook, which indicates that even with the two indicators (ref. to the screenshots), it's still hard for people to find the original notebook and determine if there's any worthy change.
\n
I don't want to start a witch hunt, instead I'd rather want everyone to go take a look at @yasufuminakama and give him an upvote if you appreciated them. Many kagglers are working hard on their notebooks and are very proud, and this unpunished plagiarism that gets rewarded with medals will only deter them to produce original work in the future.
", "rawMarkdown": "A few days ago @yasufuminakama published two great notebooks for respectively training and submit a ResNext model, which at the time achieved the highest score on the leaderboard out of the public notebooks. They are:\n* [ResNext notebook](https://www.kaggle.com/yasufuminakama/ranzcr-resnext50-32x4d-starter-training)\n* [ResNext inference notebook](https://www.kaggle.com/yasufuminakama/ranzcr-resnext50-32x4d-starter-inference)\n\nIt has been a few days now and the notebooks are sitting at ~40 upvotes each, whereas usually top notebooks in medical imaging competitions would have gotten gold by now.\n\nNow, this is understandable because there are many on-going competitions and people might not have time to compete during holidays. In fact, I wouldn't have been bothered if it wasn't for multiple public notebooks blatantly **copying @yasufuminakama's inference notebook** and changing the name so people wouldn't so easily recognize it.\n\nIn fact, one notebook that made me very angry had:\n* copied with no change, except **removing a comment linking to the original training notebook, thus elimating any reference**\n* changed the title to attract upvotes (and the title is in fact wrong)\n* stayed at the top spot for a few days\n* received over half as many upvotes as the original inference notebook\n* no response when asked about why they made it public\n\nI wouldn't be bothered by this if it was an honest mistake (e.g. accidentally made public after forking), but it is clear that the name change and erasing reference to the original notebook were not accidental or a beginner's mistake. **What disappointed me is that many of the people that upvoted the plagiarized notebook didn't even upvote the original notebook**, which indicates that even with the two indicators (ref. to the screenshots), it's still hard for people to find the original notebook and determine if there's any worthy change.\n\n\n\n\n\nI don't want to start a witch hunt, instead I'd rather want everyone to go take a look at @yasufuminakama and give him an upvote if you appreciated them. Many kagglers are working hard on their notebooks and are very proud, and this unpunished plagiarism that gets rewarded with medals will only deter them to produce original work in the future.", "votes": 141 }, { "id": 1118153, "postDate": "2020-12-18T19:30:00.030Z", "content": "Thank you xhlulu for raising this issue, I felt sad about it…
", "rawMarkdown": "Thank you xhlulu for raising this issue, I felt sad about it...", "votes": 20, "replies": [ { "id": 1118158, "postDate": "2020-12-18T19:36:40.867Z", "content": "Please keep up the good work. In most competitions people properly cite their sources and are honest about their work, I hope this is an outlier case and not the new norm.
", "rawMarkdown": "Please keep up the good work. In most competitions people properly cite their sources and are honest about their work, I hope this is an outlier case and not the new norm.", "votes": 10 } ] }, { "id": 1118532, "postDate": "2020-12-19T06:46:17.183Z", "content": "to solve the root of the problem, the upvote rules can be changed.
\ne.g. if you upvote a copied notebook, the vote goes to the original source , etc.
I feel this will solve the problem of accidentally publicizing notebooks, but might incentivize people to directly copy paste the notebook rather than using the forking option.
", "rawMarkdown": "I feel this will solve the problem of accidentally publicizing notebooks, but might incentivize people to directly copy paste the notebook rather than using the forking option.", "votes": 9 }, { "id": 1126450, "postDate": "2020-12-25T16:14:48.573Z", "content": "or if you copy a notebook it will automatically upvote the original notebook !!! because i see many notebooks being copied 100 + times but the upvote remains at 20 or 30 votes only , you should vote if you are forking the notebook !
", "rawMarkdown": "or if you copy a notebook it will automatically upvote the original notebook !!! because i see many notebooks being copied 100 + times but the upvote remains at 20 or 30 votes only , you should vote if you are forking the notebook !", "votes": 2 } ] }, { "id": 1118086, "postDate": "2020-12-18T18:04:42.143Z", "content": "This is what i like of the Kaggle community, you can't release a plagiarized notebook without consequences. Many people has reported notebooks in past competitions and at least for that competition people stop the plagiarims.
", "rawMarkdown": "This is what i like of the Kaggle community, you can't release a plagiarized notebook without consequences. Many people has reported notebooks in past competitions and at least for that competition people stop the plagiarims.", "votes": 7, "replies": [ { "id": 1118089, "postDate": "2020-12-18T18:08:06.473Z", "content": "Does kaggle remove the notebooks? I'm not familiar with the report system.
", "rawMarkdown": "Does kaggle remove the notebooks? I'm not familiar with the report system.", "votes": 2 }, { "id": 1118160, "postDate": "2020-12-18T19:38:21.400Z", "content": "It's hard to say because the notebooks are removed and i can't provide a link, therefore i don't know if the author or Kaggle remove it.
", "rawMarkdown": "It's hard to say because the notebooks are removed and i can't provide a link, therefore i don't know if the author or Kaggle remove it.", "votes": 1 }, { "id": 1118162, "postDate": "2020-12-18T19:41:30.127Z", "content": "No worries. It would be nice if it says \"removed for plagiarism\" or something so people are reminded to be careful.
", "rawMarkdown": "No worries. It would be nice if it says \"removed for plagiarism\" or something so people are reminded to be careful. " }, { "id": 1214363, "postDate": "2021-02-22T19:30:16.830Z", "content": "I think Kaggle is taking more drastic measures now with plagiarism. Ive seen a couple of profiles who were known for plagiarism and now they are not in Kaggle anymore.
", "rawMarkdown": "I think Kaggle is taking more drastic measures now with plagiarism. Ive seen a couple of profiles who were known for plagiarism and now they are not in Kaggle anymore." } ] }, { "id": 1121102, "postDate": "2020-12-21T11:26:01.380Z", "content": "Might sound harsh but name and shame is a better strategy. He/she might even brag about how they climbed up the ranking ladder with upvotes aquired from such notebooks.
", "rawMarkdown": "Might sound harsh but **name and shame** is a better strategy. He/she might even brag about how they climbed up the ranking ladder with upvotes aquired from such notebooks.", "votes": 5 }, { "id": 1124318, "postDate": "2020-12-23T20:17:15.467Z", "content": "Dear @yasufuminakama Your work is always inspirational and thought me a lot. Thanks for your endless support to the community. And @xhlulu Thanks for raising the issue. I am a beginner as well in the community but try to learn from the notebooks and upvote them as soon as I learn any single information. I don't understand why people behave like this? What's the point? Fame or being cool…?
", "rawMarkdown": "Dear @yasufuminakama Your work is always inspirational and thought me a lot. Thanks for your endless support to the community. And @xhlulu Thanks for raising the issue. I am a beginner as well in the community but try to learn from the notebooks and upvote them as soon as I learn any single information. I don't understand why people behave like this? What's the point? Fame or being cool...?", "votes": 3 }, { "id": 1119463, "postDate": "2020-12-20T05:05:27.317Z", "content": "Brother @yasufuminakama, Your notebook inspired me a lot. As I'm a beginner, it has taught me a lot. I'm following your way of approach. Thank you for your great work.
", "rawMarkdown": "Brother @yasufuminakama, Your notebook inspired me a lot. As I'm a beginner, it has taught me a lot. I'm following your way of approach. Thank you for your great work. ", "votes": 3, "replies": [ { "id": 1119853, "postDate": "2020-12-20T12:03:46.377Z", "content": "Glad to hear that :)
", "rawMarkdown": "Glad to hear that :)", "votes": 2 } ] }, { "id": 1118324, "postDate": "2020-12-19T00:20:12.330Z", "content": "Good work @yasufuminakama! I have upvoted.
\nWe should always encourage valuable sharings and avoid meaningless copying.
", "rawMarkdown": "Good work @yasufuminakama! I have upvoted.\n\nWe should always encourage valuable sharings and avoid meaningless copying.", "votes": 4, "replies": [ { "id": 1118356, "postDate": "2020-12-19T01:55:27.840Z", "content": "Exactly, thank you @gogo827jz :) !
", "rawMarkdown": "Exactly, thank you @gogo827jz :) ! ", "votes": 3 } ] }, { "id": 1214260, "postDate": "2021-02-22T17:28:49.113Z", "content": "I very genuinely thank all those \"original\" contributors spending so much of their valuable time and effort to help the community improve and from whom I've learned and been learning so much. Thank you so much!
", "rawMarkdown": "I very genuinely thank all those \"original\" contributors spending so much of their valuable time and effort to help the community improve and from whom I've learned and been learning so much. Thank you so much!", "votes": 1 }, { "id": 1131403, "postDate": "2020-12-29T17:53:09.517Z", "content": "Rather than just complaining about this, I tried to formulate some constructive suggestions in this thread. I hope the new thread might help the Kaggle team in some way even though they might not be incorporated as is.
\n", "rawMarkdown": "Rather than just complaining about this, I tried to formulate some constructive suggestions in [this thread](https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification/discussion/207431). I hope the new thread might help the Kaggle team in some way even though they might not be incorporated as is.\n\nCC: @maggiemd @philculliton ", "votes": 1 }, { "id": 1130906, "postDate": "2020-12-29T12:03:44.747Z", "content": "I'm a beginner, so I can't be sure I won't make such a mistake.
\nThanks to the reminder by a detailed and advanced person, I have been inspired to be more careful myself.
\nThank you very much.
Just created a topic in the beginners community to increase the reach of this discussion. Hopefully people take a note!
", "rawMarkdown": "Just created a [topic](https://www.kaggle.com/discussion/206639) in the beginners community to increase the reach of this discussion. Hopefully people take a note! ", "votes": 1 }, { "id": 1121340, "postDate": "2020-12-21T15:15:35.803Z", "content": "Sadly, it's not new. 🙁
\nInitially, I thought these are practices by very beginners but eventually, there are some medal suckers. It's ok to ask in the notebook to upvote if they like the notebook. But it's not ok in the above cases and also scamming the comment section to ask for upvoting of their notebook also. I saw many people reach the master tier by this, what a shame!
Kaggle should make some mechanism that would differentiate the original work and copied work. For example a blue color mark for original and yellow for copied; something like that. And also the copied notebook MUST contain the redirect source link of the original work and that would not possible to delete also in the copied notebook.
", "rawMarkdown": "Sadly, it's not new. 🙁\nInitially, I thought these are practices by **very beginners** but eventually, there are some **medal suckers**. It's **ok** to ask in the notebook to upvote if they like the notebook. But it's **not ok** in the above cases and also scamming the comment section to ask for upvoting of their notebook also. I saw many people reach the master tier by this, what a shame! \n\nKaggle should make some **mechanism** that would **differentiate** the original work and copied work. For example a blue color mark for original and yellow for copied; something like that. And also the copied notebook **MUST** contain the **redirect source link** of the **original work** and that **would not possible to delete** also in the copied notebook.", "votes": 1, "replies": [ { "id": 1121449, "postDate": "2020-12-21T16:40:49.917Z", "content": "\n\nAnd also the copied notebook MUST contain the redirect source link of the original work and that would not possible to delete also in the copied notebook.
\n
It's already there, but it's too easy to miss.
", "rawMarkdown": "> And also the copied notebook MUST contain the redirect source link of the original work and that would not possible to delete also in the copied notebook.\n\nIt's already there, but it's too easy to miss." }, { "id": 1214371, "postDate": "2021-02-22T19:39:38.040Z", "content": "Agree, however this trend started to happen more frequently I would say in the past year. When I started in Kaggle it was a much more nice environment. I also agree with the common spamming that you see in the comment section however, I think this is hard to control. For instance, if I see the author is working on a similar notebook that I am currently working and I think the author of that notebook could share interesting insights I will provide a link to a project I am currently working. My best notebooks come from constructive criticisms of others. However, when you do it constantly and spamming all the time in order for people to upvote notebooks then I think this harms the community. I dont know why people are so emphasized in getting into an upper rank. People should enjoy the learning its not like being a master or grandmaster will land you a job. Anyways, agreed with what you have just said :)
", "rawMarkdown": "Agree, however this trend started to happen more frequently I would say in the past year. When I started in Kaggle it was a much more nice environment. I also agree with the common spamming that you see in the comment section however, I think this is hard to control. For instance, if I see the author is working on a similar notebook that I am currently working and I think the author of that notebook could share interesting insights I will provide a link to a project I am currently working. My best notebooks come from constructive criticisms of others. However, when you do it constantly and spamming all the time in order for people to upvote notebooks then I think this harms the community. I dont know why people are so emphasized in getting into an upper rank. People should enjoy the learning its not like being a master or grandmaster will land you a job. Anyways, agreed with what you have just said :)" } ] }, { "id": 1131532, "postDate": "2020-12-29T19:37:26.940Z", "content": "Just found that plagiarism account had been deactivated few minutes ago.
", "rawMarkdown": "Just found that plagiarism account had been deactivated few minutes ago.", "votes": 2 }, { "id": 1123828, "postDate": "2020-12-23T14:31:29.980Z", "content": "I guess it is time that kaggle creates an equivalent of a backlink system on google. i.e. let's say I copy a notebook of 1000 lines. Then I make 10 line changes. Then f(1000/1010) of my notebook upvotes should go to the original notebook. And only f(10/1010) of the notebook should come to my account. This will therefore eliminate the incentive for plagiarism. the function I guess can be fitted from Kaggle's own dataset; by comparing notebook performance and competition performance or something similar. Would like to know your thoughts on this @xhlulu. Also, I think a similar thing should be adopted in the kaggle notebook recommendation algorithm. There is always a tradeoff between relevance and variety in a search ranking. I currently believe kaggle notebooks which are more similar cluster in the top. We can again use the individual contribution factor in the notebooks to increase the variety and therefore let more notebooks come up.
", "rawMarkdown": "I guess it is time that kaggle creates an equivalent of a backlink system on google. i.e. let's say I copy a notebook of 1000 lines. Then I make 10 line changes. Then f(1000/1010) of my notebook upvotes should go to the original notebook. And only f(10/1010) of the notebook should come to my account. This will therefore eliminate the incentive for plagiarism. the function I guess can be fitted from Kaggle's own dataset; by comparing notebook performance and competition performance or something similar. Would like to know your thoughts on this @xhlulu. Also, I think a similar thing should be adopted in the kaggle notebook recommendation algorithm. There is always a tradeoff between relevance and variety in a search ranking. I currently believe kaggle notebooks which are more similar cluster in the top. We can again use the individual contribution factor in the notebooks to increase the variety and therefore let more notebooks come up.", "votes": 2, "replies": [ { "id": 1123855, "postDate": "2020-12-23T14:46:11.510Z", "content": "I think this will push people to copy paste code manually instead of forking.
", "rawMarkdown": "I think this will push people to copy paste code manually instead of forking." }, { "id": 1123857, "postDate": "2020-12-23T14:47:15.587Z", "content": "I do think original notebooks should get a view boost through the hot algorithm.
", "rawMarkdown": "I do think original notebooks should get a view boost through the hot algorithm.", "votes": 3 } ] }, { "id": 1123614, "postDate": "2020-12-23T11:22:53.873Z", "content": "This is very serious problem. Kaggle guys should find any solution on this problem!
", "rawMarkdown": "This is very serious problem. Kaggle guys should find any solution on this problem! ", "votes": 2 }, { "id": 1121904, "postDate": "2020-12-22T02:33:50.170Z", "content": "Guess what! I have seen one plagiarist who became master recently with his copy paste notebook in this competition and was boasting on Linkedin!
", "rawMarkdown": "Guess what! I have seen one plagiarist who became master recently with his copy paste notebook in this competition and was boasting on Linkedin! ", "votes": 2, "replies": [ { "id": 1121935, "postDate": "2020-12-22T03:20:51.093Z", "content": "Pretty sure we are talking about the same person :)
", "rawMarkdown": "Pretty sure we are talking about the same person :)", "votes": 2 } ] }, { "id": 1119879, "postDate": "2020-12-20T12:26:40.020Z", "content": "Don't you think, every person who puts an effort to make a notebook public should get some credit. I spent 40 hours to make my tf2/keras notebook public but the Kaggle recommendation algorithm won't promote it for some reason. No matter whenever I come my notebook is always listed in the last.
", "rawMarkdown": "Don't you think, every person who puts an effort to make a notebook public should get some credit. I spent 40 hours to make my tf2/keras notebook public but the Kaggle recommendation algorithm won't promote it for some reason. No matter whenever I come my notebook is always listed in the last.", "votes": 2, "replies": [ { "id": 1123830, "postDate": "2020-12-23T14:32:02.973Z", "content": "sir I have suggested a system change in my comment. Would like to know your thought on this.
", "rawMarkdown": "sir I have suggested a system change in my comment. Would like to know your thought on this." } ] }, { "id": 1118568, "postDate": "2020-12-19T07:43:12.280Z", "content": "To be honest if people are not willing to make sacrifice and effort they should not even try to be in this field. I would also like to point out (maybe it does not apply to this case), that there are people who like to copy and edit your notebook, although this does not necessarily mean that they plagarize however, I will say is for them to get ideas for the notebook or competition they are working on which is totally fine. My issue is when they copy and edit a notebook and users don't give upvotes since I would assume that if you copy and edit a notebook (in old times we will call it \"Forking\" a notebook) that you actually like the notebook so why not give an upvote with just a click? On top of this, it gives me the impression that the more \"Copy and edit\" notebooks you have versus the number of upvotes you have, it will affect your overall rank (I could be wrong with this, not entirely sure). If this is the case why would copying and editing your notebook affect your rank if people are using your script to get other further ideas?
", "rawMarkdown": "To be honest if people are not willing to make sacrifice and effort they should not even try to be in this field. I would also like to point out (maybe it does not apply to this case), that there are people who like to **copy and edit** your notebook, although this does not necessarily mean that they plagarize however, I will say is for them to get ideas for the notebook or competition they are working on which is totally fine. My issue is when they copy and edit a notebook and users don't give upvotes since I would assume that if you copy and edit a notebook (in old times we will call it \"Forking\" a notebook) that you actually like the notebook so why not give an upvote with just a click? On top of this, it gives me the impression that the more \"Copy and edit\" notebooks you have versus the number of upvotes you have, it will affect your overall rank (I could be wrong with this, not entirely sure). If this is the case why would copying and editing your notebook affect your rank if people are using your script to get other further ideas?", "votes": 1 }, { "id": 1120525, "postDate": "2020-12-20T21:46:11.403Z", "content": "I've seen a similar situation in many competitions. Moreover, I saw kagglers who copied other people's notebooks one to one (without upvotes and links to original work), and at the same time were very indignant and angry when someone did the same with their copies.
\nAnd yes, these notebooks got a lot of upvotes.
Thank you for your kind words @mohamedsamirali :)
", "rawMarkdown": "Thank you for your kind words @mohamedsamirali :)", "votes": 1 } ] }, { "id": 1118071, "postDate": "2020-12-18T17:41:51.547Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 1118153, "author_name": "Y.Nakama", "author_url": "", "post_date": "2020-12-18T19:30:00.030000", "content": "Thank you xhlulu for raising this issue, I felt sad about it…
", "votes": 20, "replies": [ { "id": 1118158, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-18T19:36:40.867000", "content": "Please keep up the good work. In most competitions people properly cite their sources and are honest about their work, I hope this is an outlier case and not the new norm.
", "votes": 10, "replies": [] } ] }, { "id": 1118532, "author_name": "hengck23", "author_url": "", "post_date": "2020-12-19T06:46:17.183000", "content": "to solve the root of the problem, the upvote rules can be changed.
\ne.g. if you upvote a copied notebook, the vote goes to the original source , etc.
I feel this will solve the problem of accidentally publicizing notebooks, but might incentivize people to directly copy paste the notebook rather than using the forking option.
", "votes": 9, "replies": [] }, { "id": 1126450, "author_name": "Shubham Thapa", "author_url": "", "post_date": "2020-12-25T16:14:48.573000", "content": "or if you copy a notebook it will automatically upvote the original notebook !!! because i see many notebooks being copied 100 + times but the upvote remains at 20 or 30 votes only , you should vote if you are forking the notebook !
", "votes": 2, "replies": [] } ] }, { "id": 1118086, "author_name": "Hiram Coria 🧬", "author_url": "", "post_date": "2020-12-18T18:04:42.143000", "content": "This is what i like of the Kaggle community, you can't release a plagiarized notebook without consequences. Many people has reported notebooks in past competitions and at least for that competition people stop the plagiarims.
", "votes": 7, "replies": [ { "id": 1118089, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-18T18:08:06.473000", "content": "Does kaggle remove the notebooks? I'm not familiar with the report system.
", "votes": 2, "replies": [] }, { "id": 1118160, "author_name": "Hiram Coria 🧬", "author_url": "", "post_date": "2020-12-18T19:38:21.400000", "content": "It's hard to say because the notebooks are removed and i can't provide a link, therefore i don't know if the author or Kaggle remove it.
", "votes": 1, "replies": [] }, { "id": 1118162, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-18T19:41:30.127000", "content": "No worries. It would be nice if it says \"removed for plagiarism\" or something so people are reminded to be careful.
", "votes": 0, "replies": [] }, { "id": 1214363, "author_name": "Janio Martinez Bachmann", "author_url": "", "post_date": "2021-02-22T19:30:16.830000", "content": "I think Kaggle is taking more drastic measures now with plagiarism. Ive seen a couple of profiles who were known for plagiarism and now they are not in Kaggle anymore.
", "votes": 0, "replies": [] } ] }, { "id": 1121102, "author_name": "A Legacy Grandmaster!", "author_url": "", "post_date": "2020-12-21T11:26:01.380000", "content": "Might sound harsh but name and shame is a better strategy. He/she might even brag about how they climbed up the ranking ladder with upvotes aquired from such notebooks.
", "votes": 5, "replies": [] }, { "id": 1124318, "author_name": "Sinan Calisir", "author_url": "", "post_date": "2020-12-23T20:17:15.467000", "content": "Dear @yasufuminakama Your work is always inspirational and thought me a lot. Thanks for your endless support to the community. And @xhlulu Thanks for raising the issue. I am a beginner as well in the community but try to learn from the notebooks and upvote them as soon as I learn any single information. I don't understand why people behave like this? What's the point? Fame or being cool…?
", "votes": 3, "replies": [] }, { "id": 1119463, "author_name": "TEnsorSAge", "author_url": "", "post_date": "2020-12-20T05:05:27.317000", "content": "Brother @yasufuminakama, Your notebook inspired me a lot. As I'm a beginner, it has taught me a lot. I'm following your way of approach. Thank you for your great work.
", "votes": 3, "replies": [ { "id": 1119853, "author_name": "Y.Nakama", "author_url": "", "post_date": "2020-12-20T12:03:46.377000", "content": "Glad to hear that :)
", "votes": 2, "replies": [] } ] }, { "id": 1118324, "author_name": "Yirun Zhang", "author_url": "", "post_date": "2020-12-19T00:20:12.330000", "content": "Good work @yasufuminakama! I have upvoted.
\nWe should always encourage valuable sharings and avoid meaningless copying.
", "votes": 4, "replies": [ { "id": 1118356, "author_name": "Y.Nakama", "author_url": "", "post_date": "2020-12-19T01:55:27.840000", "content": "Exactly, thank you @gogo827jz :) !
", "votes": 3, "replies": [] } ] }, { "id": 1214260, "author_name": "Yusuf Chabuk", "author_url": "", "post_date": "2021-02-22T17:28:49.113000", "content": "I very genuinely thank all those \"original\" contributors spending so much of their valuable time and effort to help the community improve and from whom I've learned and been learning so much. Thank you so much!
", "votes": 1, "replies": [] }, { "id": 1131403, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-29T17:53:09.517000", "content": "Rather than just complaining about this, I tried to formulate some constructive suggestions in this thread. I hope the new thread might help the Kaggle team in some way even though they might not be incorporated as is.
\n", "votes": 1, "replies": [] }, { "id": 1130906, "author_name": "ANZ", "author_url": "", "post_date": "2020-12-29T12:03:44.747000", "content": "I'm a beginner, so I can't be sure I won't make such a mistake.
\nThanks to the reminder by a detailed and advanced person, I have been inspired to be more careful myself.
\nThank you very much.
Just created a topic in the beginners community to increase the reach of this discussion. Hopefully people take a note!
", "votes": 1, "replies": [] }, { "id": 1121340, "author_name": "Innat", "author_url": "", "post_date": "2020-12-21T15:15:35.803000", "content": "Sadly, it's not new. 🙁
\nInitially, I thought these are practices by very beginners but eventually, there are some medal suckers. It's ok to ask in the notebook to upvote if they like the notebook. But it's not ok in the above cases and also scamming the comment section to ask for upvoting of their notebook also. I saw many people reach the master tier by this, what a shame!
Kaggle should make some mechanism that would differentiate the original work and copied work. For example a blue color mark for original and yellow for copied; something like that. And also the copied notebook MUST contain the redirect source link of the original work and that would not possible to delete also in the copied notebook.
", "votes": 1, "replies": [ { "id": 1121449, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-21T16:40:49.917000", "content": "\n\nAnd also the copied notebook MUST contain the redirect source link of the original work and that would not possible to delete also in the copied notebook.
\n
It's already there, but it's too easy to miss.
", "votes": 0, "replies": [] }, { "id": 1214371, "author_name": "Janio Martinez Bachmann", "author_url": "", "post_date": "2021-02-22T19:39:38.040000", "content": "Agree, however this trend started to happen more frequently I would say in the past year. When I started in Kaggle it was a much more nice environment. I also agree with the common spamming that you see in the comment section however, I think this is hard to control. For instance, if I see the author is working on a similar notebook that I am currently working and I think the author of that notebook could share interesting insights I will provide a link to a project I am currently working. My best notebooks come from constructive criticisms of others. However, when you do it constantly and spamming all the time in order for people to upvote notebooks then I think this harms the community. I dont know why people are so emphasized in getting into an upper rank. People should enjoy the learning its not like being a master or grandmaster will land you a job. Anyways, agreed with what you have just said :)
", "votes": 0, "replies": [] } ] }, { "id": 1131532, "author_name": "Chew Kok Wah", "author_url": "", "post_date": "2020-12-29T19:37:26.940000", "content": "Just found that plagiarism account had been deactivated few minutes ago.
", "votes": 2, "replies": [] }, { "id": 1123828, "author_name": "shyambhu Mukherjee", "author_url": "", "post_date": "2020-12-23T14:31:29.980000", "content": "I guess it is time that kaggle creates an equivalent of a backlink system on google. i.e. let's say I copy a notebook of 1000 lines. Then I make 10 line changes. Then f(1000/1010) of my notebook upvotes should go to the original notebook. And only f(10/1010) of the notebook should come to my account. This will therefore eliminate the incentive for plagiarism. the function I guess can be fitted from Kaggle's own dataset; by comparing notebook performance and competition performance or something similar. Would like to know your thoughts on this @xhlulu. Also, I think a similar thing should be adopted in the kaggle notebook recommendation algorithm. There is always a tradeoff between relevance and variety in a search ranking. I currently believe kaggle notebooks which are more similar cluster in the top. We can again use the individual contribution factor in the notebooks to increase the variety and therefore let more notebooks come up.
", "votes": 2, "replies": [ { "id": 1123855, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-23T14:46:11.510000", "content": "I think this will push people to copy paste code manually instead of forking.
", "votes": 0, "replies": [] }, { "id": 1123857, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-23T14:47:15.587000", "content": "I do think original notebooks should get a view boost through the hot algorithm.
", "votes": 3, "replies": [] } ] }, { "id": 1123614, "author_name": "Saurabh Shahane", "author_url": "", "post_date": "2020-12-23T11:22:53.873000", "content": "This is very serious problem. Kaggle guys should find any solution on this problem!
", "votes": 2, "replies": [] }, { "id": 1121904, "author_name": "Sina Mehdinia", "author_url": "", "post_date": "2020-12-22T02:33:50.170000", "content": "Guess what! I have seen one plagiarist who became master recently with his copy paste notebook in this competition and was boasting on Linkedin!
", "votes": 2, "replies": [ { "id": 1121935, "author_name": "xhlulu", "author_url": "", "post_date": "2020-12-22T03:20:51.093000", "content": "Pretty sure we are talking about the same person :)
", "votes": 2, "replies": [] } ] }, { "id": 1119879, "author_name": "Harveen Singh Chadha", "author_url": "", "post_date": "2020-12-20T12:26:40.020000", "content": "Don't you think, every person who puts an effort to make a notebook public should get some credit. I spent 40 hours to make my tf2/keras notebook public but the Kaggle recommendation algorithm won't promote it for some reason. No matter whenever I come my notebook is always listed in the last.
", "votes": 2, "replies": [ { "id": 1123830, "author_name": "shyambhu Mukherjee", "author_url": "", "post_date": "2020-12-23T14:32:02.973000", "content": "sir I have suggested a system change in my comment. Would like to know your thought on this.
", "votes": 0, "replies": [] } ] }, { "id": 1118568, "author_name": "Janio Martinez Bachmann", "author_url": "", "post_date": "2020-12-19T07:43:12.280000", "content": "To be honest if people are not willing to make sacrifice and effort they should not even try to be in this field. I would also like to point out (maybe it does not apply to this case), that there are people who like to copy and edit your notebook, although this does not necessarily mean that they plagarize however, I will say is for them to get ideas for the notebook or competition they are working on which is totally fine. My issue is when they copy and edit a notebook and users don't give upvotes since I would assume that if you copy and edit a notebook (in old times we will call it \"Forking\" a notebook) that you actually like the notebook so why not give an upvote with just a click? On top of this, it gives me the impression that the more \"Copy and edit\" notebooks you have versus the number of upvotes you have, it will affect your overall rank (I could be wrong with this, not entirely sure). If this is the case why would copying and editing your notebook affect your rank if people are using your script to get other further ideas?
", "votes": 1, "replies": [] }, { "id": 1120525, "author_name": "Maksym Shkliarevskyi", "author_url": "", "post_date": "2020-12-20T21:46:11.403000", "content": "I've seen a similar situation in many competitions. Moreover, I saw kagglers who copied other people's notebooks one to one (without upvotes and links to original work), and at the same time were very indignant and angry when someone did the same with their copies.
\nAnd yes, these notebooks got a lot of upvotes.
Thank you for your kind words @mohamedsamirali :)
", "votes": 1, "replies": [] } ] }, { "id": 1118071, "author_name": "", "author_url": "", "post_date": "2020-12-18T17:41:51.547000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1118070": "A few days ago @yasufuminakama published two great notebooks for respectively training and submit a ResNext model, which at the time achieved the highest score on the leaderboard out of the public notebooks. They are:\n* [ResNext notebook](https://www.kaggle.com/yasufuminakama/ranzcr-resnext50-32x4d-starter-training)\n* [ResNext inference notebook](https://www.kaggle.com/yasufuminakama/ranzcr-resnext50-32x4d-starter-inference)\n\nIt has been a few days now and the notebooks are sitting at ~40 upvotes each, whereas usually top notebooks in medical imaging competitions would have gotten gold by now.\n\nNow, this is understandable because there are many on-going competitions and people might not have time to compete during holidays. In fact, I wouldn't have been bothered if it wasn't for multiple public notebooks blatantly **copying @yasufuminakama's inference notebook** and changing the name so people wouldn't so easily recognize it.\n\nIn fact, one notebook that made me very angry had:\n* copied with no change, except **removing a comment linking to the original training notebook, thus elimating any reference**\n* changed the title to attract upvotes (and the title is in fact wrong)\n* stayed at the top spot for a few days\n* received over half as many upvotes as the original inference notebook\n* no response when asked about why they made it public\n\nI wouldn't be bothered by this if it was an honest mistake (e.g. accidentally made public after forking), but it is clear that the name change and erasing reference to the original notebook were not accidental or a beginner's mistake. **What disappointed me is that many of the people that upvoted the plagiarized notebook didn't even upvote the original notebook**, which indicates that even with the two indicators (ref. to the screenshots), it's still hard for people to find the original notebook and determine if there's any worthy change.\n\n\n\n\n\nI don't want to start a witch hunt, instead I'd rather want everyone to go take a look at @yasufuminakama and give him an upvote if you appreciated them. Many kagglers are working hard on their notebooks and are very proud, and this unpunished plagiarism that gets rewarded with medals will only deter them to produce original work in the future.", "1118153": "Thank you xhlulu for raising this issue, I felt sad about it...", "1118532": "to solve the root of the problem, the upvote rules can be changed. \ne.g. if you upvote a copied notebook, the vote goes to the original source , etc.", "1118086": "This is what i like of the Kaggle community, you can't release a plagiarized notebook without consequences. Many people has reported notebooks in past competitions and at least for that competition people stop the plagiarims.", "1121102": "Might sound harsh but **name and shame** is a better strategy. He/she might even brag about how they climbed up the ranking ladder with upvotes aquired from such notebooks.", "1124318": "Dear @yasufuminakama Your work is always inspirational and thought me a lot. Thanks for your endless support to the community. And @xhlulu Thanks for raising the issue. I am a beginner as well in the community but try to learn from the notebooks and upvote them as soon as I learn any single information. I don't understand why people behave like this? What's the point? Fame or being cool...?", "1119463": "Brother @yasufuminakama, Your notebook inspired me a lot. As I'm a beginner, it has taught me a lot. I'm following your way of approach. Thank you for your great work. ", "1118324": "Good work @yasufuminakama! I have upvoted.\n\nWe should always encourage valuable sharings and avoid meaningless copying.", "1214260": "I very genuinely thank all those \"original\" contributors spending so much of their valuable time and effort to help the community improve and from whom I've learned and been learning so much. Thank you so much!", "1131403": "Rather than just complaining about this, I tried to formulate some constructive suggestions in [this thread](https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification/discussion/207431). I hope the new thread might help the Kaggle team in some way even though they might not be incorporated as is.\n\nCC: @maggiemd @philculliton ", "1130906": "I'm a beginner, so I can't be sure I won't make such a mistake.\nThanks to the reminder by a detailed and advanced person, I have been inspired to be more careful myself.\nThank you very much.", "1126485": "Just created a [topic](https://www.kaggle.com/discussion/206639) in the beginners community to increase the reach of this discussion. Hopefully people take a note! ", "1121340": "Sadly, it's not new. 🙁\nInitially, I thought these are practices by **very beginners** but eventually, there are some **medal suckers**. It's **ok** to ask in the notebook to upvote if they like the notebook. But it's **not ok** in the above cases and also scamming the comment section to ask for upvoting of their notebook also. I saw many people reach the master tier by this, what a shame! \n\nKaggle should make some **mechanism** that would **differentiate** the original work and copied work. For example a blue color mark for original and yellow for copied; something like that. And also the copied notebook **MUST** contain the **redirect source link** of the **original work** and that **would not possible to delete** also in the copied notebook.", "1131532": "Just found that plagiarism account had been deactivated few minutes ago.", "1123828": "I guess it is time that kaggle creates an equivalent of a backlink system on google. i.e. let's say I copy a notebook of 1000 lines. Then I make 10 line changes. Then f(1000/1010) of my notebook upvotes should go to the original notebook. And only f(10/1010) of the notebook should come to my account. This will therefore eliminate the incentive for plagiarism. the function I guess can be fitted from Kaggle's own dataset; by comparing notebook performance and competition performance or something similar. Would like to know your thoughts on this @xhlulu. Also, I think a similar thing should be adopted in the kaggle notebook recommendation algorithm. There is always a tradeoff between relevance and variety in a search ranking. I currently believe kaggle notebooks which are more similar cluster in the top. We can again use the individual contribution factor in the notebooks to increase the variety and therefore let more notebooks come up.", "1123614": "This is very serious problem. Kaggle guys should find any solution on this problem! ", "1121904": "Guess what! I have seen one plagiarist who became master recently with his copy paste notebook in this competition and was boasting on Linkedin! ", "1119879": "Don't you think, every person who puts an effort to make a notebook public should get some credit. I spent 40 hours to make my tf2/keras notebook public but the Kaggle recommendation algorithm won't promote it for some reason. No matter whenever I come my notebook is always listed in the last.", "1118568": "To be honest if people are not willing to make sacrifice and effort they should not even try to be in this field. I would also like to point out (maybe it does not apply to this case), that there are people who like to **copy and edit** your notebook, although this does not necessarily mean that they plagarize however, I will say is for them to get ideas for the notebook or competition they are working on which is totally fine. My issue is when they copy and edit a notebook and users don't give upvotes since I would assume that if you copy and edit a notebook (in old times we will call it \"Forking\" a notebook) that you actually like the notebook so why not give an upvote with just a click? On top of this, it gives me the impression that the more \"Copy and edit\" notebooks you have versus the number of upvotes you have, it will affect your overall rank (I could be wrong with this, not entirely sure). If this is the case why would copying and editing your notebook affect your rank if people are using your script to get other further ideas?", "1120525": "I've seen a similar situation in many competitions. Moreover, I saw kagglers who copied other people's notebooks one to one (without upvotes and links to original work), and at the same time were very indignant and angry when someone did the same with their copies. \nAnd yes, these notebooks got a lot of upvotes.", "1119124": "", "1118071": "" } }