I have noticed it in recent kernels that people are posting every minute improvement they are making in their scores. Not that I mind but it kind of kills the competitive spirit. Please share the hints or new ideas that you're applying in your model which you think can help people and someone may learn from your approach because nobody is learning from just a forked kernel with little bit tuning instead it's crowding the leaderboard. I hope you understand the issue and refrain from sharing the direct solutions.
", "messages": [ { "id": 436317, "postDate": "2018-12-10T05:49:10.297Z", "content": "To be frank. People who are experts, masters and perhaps even Grandmasters learn every day by sharing solutions. What good comes if you keep something only for yourself. Logically speaking if certain bacteria stop sharing their by products neither you or me can live or possibly digest food.
\n\nEmbrace the competition. Keep your sportive spiriit high. And, keep sharing!
", "rawMarkdown": "To be frank. People who are experts, masters and perhaps even Grandmasters learn every day by sharing solutions. What good comes if you keep something only for yourself. Logically speaking if certain bacteria stop sharing their by products neither you or me can live or possibly digest food. \n\nEmbrace the competition. Keep your sportive spiriit high. And, keep sharing!", "votes": 12, "replies": [ { "id": 436323, "postDate": "2018-12-10T06:00:56.910Z", "content": "Competely agree. One new little idea which will improve a score of the kernel a bit, could help a huge number of people here! And then these people could improve the idea and continue the positive cycle of knowledge.
", "rawMarkdown": "Competely agree. One new little idea which will improve a score of the kernel a bit, could help a huge number of people here! And then these people could improve the idea and continue the positive cycle of knowledge.", "votes": 2 } ] }, { "id": 437812, "postDate": "2018-12-12T14:38:50.883Z", "content": "I agree on the fact that sharing solutions will lead to crowding on the leader board. However if the kernels are not shared, it would defeat the purpose of Kaggle : i.e. Learning. I also believe that instead of forking the kernel, People should take the ideas and build their kernels on them, this will not only lead to ambiguity to the solutions, but will promote learning for the beginners at different levels.
", "rawMarkdown": "I agree on the fact that sharing solutions will lead to crowding on the leader board. However if the kernels are not shared, it would defeat the purpose of Kaggle : i.e. Learning. I also believe that instead of forking the kernel, People should take the ideas and build their kernels on them, this will not only lead to ambiguity to the solutions, but will promote learning for the beginners at different levels. ", "votes": 7 }, { "id": 435045, "postDate": "2018-12-07T11:37:38.910Z", "content": "I have noticed it in recent kernels that people are posting every minute improvement they are making in their scores. Not that I mind but it kind of kills the competitive spirit. Please share the hints or new ideas that you're applying in your model which you think can help people and someone may learn from your approach because nobody is learning from just a forked kernel with little bit tuning instead it's crowding the leaderboard. I hope you understand the issue and refrain from sharing the direct solutions.
", "rawMarkdown": "I have noticed it in recent kernels that people are posting every minute improvement they are making in their scores. Not that I mind but it kind of kills the competitive spirit. Please share the hints or new ideas that you're applying in your model which you think can help people and someone may learn from your approach because nobody is learning from just a forked kernel with little bit tuning instead it's crowding the leaderboard. I hope you understand the issue and refrain from sharing the direct solutions.", "votes": 3 }, { "id": 435647, "postDate": "2018-12-08T13:04:28.257Z", "content": "the funny thing is that you are only seeing the solutions which people are posting, which is not all the solutions out there...some people dumb answers down and then snipe ;)
", "rawMarkdown": "the funny thing is that you are only seeing the solutions which people are posting, which is not all the solutions out there...some people dumb answers down and then snipe ;)", "votes": 3, "replies": [ { "id": 435649, "postDate": "2018-12-08T13:08:33.490Z", "content": "Yeah, I get that totally I didn't say I mind them posting the solutions what I meant is it's kind of not cool. And there is a long way to go in this competition yet so let's see what happens.
", "rawMarkdown": "Yeah, I get that totally I didn't say I mind them posting the solutions what I meant is it's kind of not cool. And there is a long way to go in this competition yet so let's see what happens." } ] }, { "id": 435763, "postDate": "2018-12-08T18:11:51.067Z", "content": "There are still two months till the deadline! There is plenty of time for improvement :)
", "rawMarkdown": "There are still two months till the deadline! There is plenty of time for improvement :)", "votes": 2, "replies": [ { "id": 435775, "postDate": "2018-12-08T18:45:26.923Z", "content": "Indeed, I am not worried about that at all.
", "rawMarkdown": "Indeed, I am not worried about that at all." } ] }, { "id": 435256, "postDate": "2018-12-07T19:00:59.323Z", "content": "can't agree more!
", "rawMarkdown": "can't agree more!", "votes": 2 }, { "id": 435797, "postDate": "2018-12-08T19:42:28.013Z", "content": "agreed
", "rawMarkdown": "agreed", "votes": 1 }, { "id": 435780, "postDate": "2018-12-08T18:53:24.040Z", "content": "Sharing is important and fine but sharing your best solution is not something I would suggest. Plus no one gets a gold by using a public kernel.
", "rawMarkdown": "Sharing is important and fine but sharing your best solution is not something I would suggest. Plus no one gets a gold by using a public kernel.", "votes": 1 }, { "id": 435748, "postDate": "2018-12-08T17:16:37.997Z", "content": "Agreed. But, we will be crowding the learderboard regardless. Without the option of exploring undiscovered data set, we, usually, will come down to models with similar performances. A breakthrough, like CNN, would rarely happen.
", "rawMarkdown": "Agreed. But, we will be crowding the learderboard regardless. Without the option of exploring undiscovered data set, we, usually, will come down to models with similar performances. A breakthrough, like CNN, would rarely happen.", "votes": 1 }, { "id": 435511, "postDate": "2018-12-08T06:17:08.797Z", "content": "Me too, no longer care about the leaderboard anymore. Tuning the model to such an extent (mostly from existing Light GBM kernels) is not as substantial as discovering new insights from the data by perusing it carefully.
", "rawMarkdown": "Me too, no longer care about the leaderboard anymore. Tuning the model to such an extent (mostly from existing Light GBM kernels) is not as substantial as discovering new insights from the data by perusing it carefully.", "votes": 2 }, { "id": 436121, "postDate": "2018-12-09T16:33:01.993Z", "content": "I don't think Kaggle is highly competitive place.\nIn my opinion Kaggle is where I can learn through other's kernels and experiment it.\nBut also I do think that by sharing the kernels it push the competition to a higher results too.
", "rawMarkdown": "I don't think Kaggle is highly competitive place.\nIn my opinion Kaggle is where I can learn through other's kernels and experiment it.\nBut also I do think that by sharing the kernels it push the competition to a higher results too." }, { "id": 435283, "postDate": "2018-12-07T19:41:00.067Z", "content": "Yeah I agree.. but on a purely competitive level, I don't know if it'll make a difference, at least to the top 100 or so, once the competition settles down and only the serious people remain interested.
", "rawMarkdown": "Yeah I agree.. but on a purely competitive level, I don't know if it'll make a difference, at least to the top 100 or so, once the competition settles down and only the serious people remain interested. ", "replies": [ { "id": 435484, "postDate": "2018-12-08T04:54:26.007Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ] } ], "comments": [ { "id": 436317, "author_name": "Shahebaz Mohammad", "author_url": "", "post_date": "2018-12-10T05:49:10.297000", "content": "To be frank. People who are experts, masters and perhaps even Grandmasters learn every day by sharing solutions. What good comes if you keep something only for yourself. Logically speaking if certain bacteria stop sharing their by products neither you or me can live or possibly digest food.
\n\nEmbrace the competition. Keep your sportive spiriit high. And, keep sharing!
", "votes": 12, "replies": [ { "id": 436323, "author_name": "Andrey Lukyanenko", "author_url": "", "post_date": "2018-12-10T06:00:56.910000", "content": "Competely agree. One new little idea which will improve a score of the kernel a bit, could help a huge number of people here! And then these people could improve the idea and continue the positive cycle of knowledge.
", "votes": 2, "replies": [] } ] }, { "id": 437812, "author_name": "Ashish Gupta", "author_url": "", "post_date": "2018-12-12T14:38:50.883000", "content": "I agree on the fact that sharing solutions will lead to crowding on the leader board. However if the kernels are not shared, it would defeat the purpose of Kaggle : i.e. Learning. I also believe that instead of forking the kernel, People should take the ideas and build their kernels on them, this will not only lead to ambiguity to the solutions, but will promote learning for the beginners at different levels.
", "votes": 7, "replies": [] }, { "id": 435647, "author_name": "anonemaus", "author_url": "", "post_date": "2018-12-08T13:04:28.257000", "content": "the funny thing is that you are only seeing the solutions which people are posting, which is not all the solutions out there...some people dumb answers down and then snipe ;)
", "votes": 3, "replies": [ { "id": 435649, "author_name": "Ashirwad", "author_url": "", "post_date": "2018-12-08T13:08:33.490000", "content": "Yeah, I get that totally I didn't say I mind them posting the solutions what I meant is it's kind of not cool. And there is a long way to go in this competition yet so let's see what happens.
", "votes": 0, "replies": [] } ] }, { "id": 435763, "author_name": "Andrey Lukyanenko", "author_url": "", "post_date": "2018-12-08T18:11:51.067000", "content": "There are still two months till the deadline! There is plenty of time for improvement :)
", "votes": 2, "replies": [ { "id": 435775, "author_name": "Ashirwad", "author_url": "", "post_date": "2018-12-08T18:45:26.923000", "content": "Indeed, I am not worried about that at all.
", "votes": 0, "replies": [] } ] }, { "id": 435256, "author_name": "Firat Gonen", "author_url": "", "post_date": "2018-12-07T19:00:59.323000", "content": "can't agree more!
", "votes": 2, "replies": [] }, { "id": 435797, "author_name": "Aniruddha", "author_url": "", "post_date": "2018-12-08T19:42:28.013000", "content": "agreed
", "votes": 1, "replies": [] }, { "id": 435780, "author_name": "Siddharth Yadav", "author_url": "", "post_date": "2018-12-08T18:53:24.040000", "content": "Sharing is important and fine but sharing your best solution is not something I would suggest. Plus no one gets a gold by using a public kernel.
", "votes": 1, "replies": [] }, { "id": 435748, "author_name": "Elliot", "author_url": "", "post_date": "2018-12-08T17:16:37.997000", "content": "Agreed. But, we will be crowding the learderboard regardless. Without the option of exploring undiscovered data set, we, usually, will come down to models with similar performances. A breakthrough, like CNN, would rarely happen.
", "votes": 1, "replies": [] }, { "id": 435511, "author_name": "Eddo Putradipura", "author_url": "", "post_date": "2018-12-08T06:17:08.797000", "content": "Me too, no longer care about the leaderboard anymore. Tuning the model to such an extent (mostly from existing Light GBM kernels) is not as substantial as discovering new insights from the data by perusing it carefully.
", "votes": 2, "replies": [] }, { "id": 436121, "author_name": "QnKhuat", "author_url": "", "post_date": "2018-12-09T16:33:01.993000", "content": "I don't think Kaggle is highly competitive place.\nIn my opinion Kaggle is where I can learn through other's kernels and experiment it.\nBut also I do think that by sharing the kernels it push the competition to a higher results too.
", "votes": 0, "replies": [] }, { "id": 435283, "author_name": "Panchajanya Banerjee (Pancham)", "author_url": "", "post_date": "2018-12-07T19:41:00.067000", "content": "Yeah I agree.. but on a purely competitive level, I don't know if it'll make a difference, at least to the top 100 or so, once the competition settles down and only the serious people remain interested.
", "votes": 0, "replies": [ { "id": 435484, "author_name": "", "author_url": "", "post_date": "2018-12-08T04:54:26.007000", "content": "", "votes": -1, "replies": [] } ] } ], "raw_markdown_by_id": { "436317": "To be frank. People who are experts, masters and perhaps even Grandmasters learn every day by sharing solutions. What good comes if you keep something only for yourself. Logically speaking if certain bacteria stop sharing their by products neither you or me can live or possibly digest food. \n\nEmbrace the competition. Keep your sportive spiriit high. And, keep sharing!", "437812": "I agree on the fact that sharing solutions will lead to crowding on the leader board. However if the kernels are not shared, it would defeat the purpose of Kaggle : i.e. Learning. I also believe that instead of forking the kernel, People should take the ideas and build their kernels on them, this will not only lead to ambiguity to the solutions, but will promote learning for the beginners at different levels. ", "435045": "I have noticed it in recent kernels that people are posting every minute improvement they are making in their scores. Not that I mind but it kind of kills the competitive spirit. Please share the hints or new ideas that you're applying in your model which you think can help people and someone may learn from your approach because nobody is learning from just a forked kernel with little bit tuning instead it's crowding the leaderboard. I hope you understand the issue and refrain from sharing the direct solutions.", "435647": "the funny thing is that you are only seeing the solutions which people are posting, which is not all the solutions out there...some people dumb answers down and then snipe ;)", "435763": "There are still two months till the deadline! There is plenty of time for improvement :)", "435256": "can't agree more!", "435797": "agreed", "435780": "Sharing is important and fine but sharing your best solution is not something I would suggest. Plus no one gets a gold by using a public kernel.", "435748": "Agreed. But, we will be crowding the learderboard regardless. Without the option of exploring undiscovered data set, we, usually, will come down to models with similar performances. A breakthrough, like CNN, would rarely happen.", "435511": "Me too, no longer care about the leaderboard anymore. Tuning the model to such an extent (mostly from existing Light GBM kernels) is not as substantial as discovering new insights from the data by perusing it carefully.", "436121": "I don't think Kaggle is highly competitive place.\nIn my opinion Kaggle is where I can learn through other's kernels and experiment it.\nBut also I do think that by sharing the kernels it push the competition to a higher results too.", "435283": "Yeah I agree.. but on a purely competitive level, I don't know if it'll make a difference, at least to the top 100 or so, once the competition settles down and only the serious people remain interested. " } }