I'm new to kaggle competition and very curious some people are sharing the solutions competitive with the bronze ones in public LB before the deadline.\nAre there advantages to them or the competition?
", "messages": [ { "id": 268760, "postDate": "2018-01-15T14:06:02.763Z", "content": "There can only be few winners but everyone can move forward. My target is to move forward and not about winning.
\n\nThink of winning in future, not now.
", "rawMarkdown": "There can only be few winners but everyone can move forward. My target is to move forward and not about winning. \n\nThink of winning in future, not now.", "votes": 27, "replies": [ { "id": 268787, "postDate": "2018-01-15T16:27:23.560Z", "rawMarkdown": "", "votes": -2, "isDeleted": true }, { "id": 268961, "postDate": "2018-01-16T01:36:57.573Z", "content": "Thank you for making this competition more impressive and meaningful to me.
", "rawMarkdown": "Thank you for making this competition more impressive and meaningful to me." }, { "id": 269002, "postDate": "2018-01-16T03:28:40.407Z", "content": "@CherKeng\nThank you for sharing the parts of your solutions in the forum. \nI got your idea.
", "rawMarkdown": "@CherKeng\nThank you for sharing the parts of your solutions in the forum. \nI got your idea." }, { "id": 269057, "postDate": "2018-01-16T05:47:46.327Z", "content": "@choker\nEdit:\nWhy did you post such a comment?\nIt's possible to sound uncomfortable.
", "rawMarkdown": "@choker\nEdit:\nWhy did you post such a comment?\nIt's possible to sound uncomfortable.\n" }, { "id": 269122, "postDate": "2018-01-16T08:42:59.323Z", "content": "It's much more meaningful then yours.
", "rawMarkdown": "It's much more meaningful then yours." }, { "id": 269451, "postDate": "2018-01-16T21:05:50.053Z", "content": "I have to say this is my first \"real\" Kaggle competition. I don't know any technique like ensembling, feature selection, and etc. I do learning a lot from your shared model. Without you, the new beginners hard to get those techniques or tricks that commonly used in Kaggle competitions.
\n\nThose who feel it is unfair to people working hard on their own model are compassionate. If they know ensembling, even they don't want to learn the knowledge of the shared model, they should still gain benefits from it. For the detail of ensembling, please check https://www.kaggle.com/wiki/Ensembling
\n\nUpdated: \nPublic sharing is what the Rules supported: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/rules
", "rawMarkdown": "I have to say this is my first \"real\" Kaggle competition. I don't know any technique like ensembling, feature selection, and etc. I do learning a lot from your shared model. Without you, the new beginners hard to get those techniques or tricks that commonly used in Kaggle competitions.\n\nThose who feel it is unfair to people working hard on their own model are compassionate. If they know ensembling, even they don't want to learn the knowledge of the shared model, they should still gain benefits from it. For the detail of ensembling, please check https://www.kaggle.com/wiki/Ensembling \n\nUpdated: \nPublic sharing is what the Rules supported: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/rules", "votes": 1 } ] }, { "id": 268913, "postDate": "2018-01-15T23:47:29.187Z", "content": "Also mind the fact that 99% of the organizers who host a competition benefit more from discussion and ideas shared in the forums than from the actual top solution that might never end up in their production line. That's why some competitions as this one prevent us from sharing outside from the kaggle forums. Not to mention at all as other have said that here sharing and learning is what makes this community unique. Without that kaggle would be nothing.
", "rawMarkdown": "Also mind the fact that 99% of the organizers who host a competition benefit more from discussion and ideas shared in the forums than from the actual top solution that might never end up in their production line. That's why some competitions as this one prevent us from sharing outside from the kaggle forums. Not to mention at all as other have said that here sharing and learning is what makes this community unique. Without that kaggle would be nothing.", "votes": 6, "replies": [ { "id": 269060, "postDate": "2018-01-16T05:51:11.610Z", "content": "That's an interesting fact for me.
", "rawMarkdown": "That's an interesting fact for me.", "votes": 1 } ] }, { "id": 268857, "postDate": "2018-01-15T20:33:57.487Z", "content": "i think it is pretty unfair towards people who try hard the last 3 months by themselves to reach top 10% and then someone shares recepies so everybody who has time in the last 2 weeks can easily reach 85%
", "rawMarkdown": "i think it is pretty unfair towards people who try hard the last 3 months by themselves to reach top 10% and then someone shares recepies so everybody who has time in the last 2 weeks can easily reach 85% ", "votes": 4, "replies": [ { "id": 268862, "postDate": "2018-01-15T20:45:40.467Z", "content": "Are you here for trying harder or learning better?
", "rawMarkdown": "Are you here for trying harder or learning better? ", "votes": 7 }, { "id": 269435, "postDate": "2018-01-16T20:46:49.067Z", "content": "@kirk has fair point, even though it pains me that the exact code was shared. But then it make sense if ppl can build on that. Next time I would have to watch discussion more closely.
", "rawMarkdown": "@kirk has fair point, even though it pains me that the exact code was shared. But then it make sense if ppl can build on that. Next time I would have to watch discussion more closely." }, { "id": 269506, "postDate": "2018-01-16T23:55:04.580Z", "content": "I've got another comment on email. And I have to say: almost nobody learns nothing after competition end. I don't know why is this but somehow competition setting makes you really want to learn and try thing to compete. It is clear that looking at previous competitions: there is no activity, nobody spends fulltime effort, nobody creates new techniques, nobody even repeats previous solutions for the most part. \nAnd ability to mess with leaderboard by disclosing good enough solutions is the part of fun being top performer here ^_^. Learn fast, do things, be first, don't care about what others doing.
", "rawMarkdown": "I've got another comment on email. And I have to say: almost nobody learns nothing after competition end. I don't know why is this but somehow competition setting makes you really want to learn and try thing to compete. It is clear that looking at previous competitions: there is no activity, nobody spends fulltime effort, nobody creates new techniques, nobody even repeats previous solutions for the most part. \nAnd ability to mess with leaderboard by disclosing good enough solutions is the part of fun being top performer here ^_^. Learn fast, do things, be first, don't care about what others doing.", "votes": 6 }, { "id": 269513, "postDate": "2018-01-17T00:13:42.360Z", "content": "@Sergey - When I was competing, after every competition, I vowed to learn from the top solutions. But, in reality, I always just gave them a quick glance and then jumped into the next active competition. So, I totally agree with you!
", "rawMarkdown": "@Sergey - When I was competing, after every competition, I _vowed_ to learn from the top solutions. But, in reality, I always just gave them a quick glance and then jumped into the next active competition. So, I totally agree with you!", "votes": 5 }, { "id": 269542, "postDate": "2018-01-17T01:13:08.513Z", "content": "@Sergey Mushinskiy\nI understood it is important to share before the deadline by your post.\nThank you.
", "rawMarkdown": "@Sergey Mushinskiy\nI understood it is important to share before the deadline by your post.\nThank you.", "votes": 1 } ] }, { "id": 268742, "postDate": "2018-01-15T13:13:04.023Z", "content": "Solution sharing may lead to more advanced solutions. I.e. if everybody shares different solutions reasonably far from competition deadline it make possible to develop new, original approaches on top of them.
", "rawMarkdown": "Solution sharing may lead to more advanced solutions. I.e. if everybody shares different solutions reasonably far from competition deadline it make possible to develop new, original approaches on top of them. ", "votes": 4 }, { "id": 268720, "postDate": "2018-01-15T11:42:38.207Z", "content": "I'm new to kaggle competition and very curious some people are sharing the solutions competitive with the bronze ones in public LB before the deadline.\nAre there advantages to them or the competition?
", "rawMarkdown": "I'm new to kaggle competition and very curious some people are sharing the solutions competitive with the bronze ones in public LB **before the deadline.**\nAre there advantages to them or the competition?", "votes": 2 }, { "id": 269852, "postDate": "2018-01-17T12:09:03.233Z", "content": "@kakapo2
\n\nyou can also take a look at this, some people offer tips and solution for another kaggle competition. I learn a lot from them:
\n\nhttps://www.kaggle.com/c/statoil-iceberg-classifier-challenge/discussion/47062
\n\nin some case, even csv files are provided for experiments in ensemble:
\n\nhttps://www.kaggle.com/dongxu027/explore-stacking-lb-0-1463
", "rawMarkdown": "@kakapo2\n\nyou can also take a look at this, some people offer tips and solution for another kaggle competition. I learn a lot from them:\n\nhttps://www.kaggle.com/c/statoil-iceberg-classifier-challenge/discussion/47062\n\nin some case, even csv files are provided for experiments in ensemble:\n\n\nhttps://www.kaggle.com/dongxu027/explore-stacking-lb-0-1463\n", "votes": 1, "replies": [ { "id": 270245, "postDate": "2018-01-18T00:21:39.803Z", "content": "Thank you CherKeng, I'll read them.
", "rawMarkdown": "Thank you CherKeng, I'll read them." } ] }, { "id": 269056, "postDate": "2018-01-16T05:47:08.667Z", "content": "Thank you all for answering my question.
\nOf course, I know it's a silly question, especially for experienced kagglers.
\nHowever, I believe it helps I or some new kagglers could understand that a competition in kaggle is not only a \"competition\".
Strategy Deception
", "rawMarkdown": "Strategy Deception", "votes": 2, "replies": [ { "id": 269062, "postDate": "2018-01-16T05:54:02.337Z", "content": "Oh, do you mean some of posted solutions are aimed to mislead other competitiors?
", "rawMarkdown": "Oh, do you mean some of posted solutions are aimed to mislead other competitiors?" } ] } ], "comments": [ { "id": 268760, "author_name": "hengck23", "author_url": "", "post_date": "2018-01-15T14:06:02.763000", "content": "There can only be few winners but everyone can move forward. My target is to move forward and not about winning.
\n\nThink of winning in future, not now.
", "votes": 27, "replies": [ { "id": 268787, "author_name": "", "author_url": "", "post_date": "2018-01-15T16:27:23.560000", "content": "", "votes": -2, "replies": [] }, { "id": 268961, "author_name": "Yibing Wu", "author_url": "", "post_date": "2018-01-16T01:36:57.573000", "content": "Thank you for making this competition more impressive and meaningful to me.
", "votes": 0, "replies": [] }, { "id": 269002, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-16T03:28:40.407000", "content": "@CherKeng\nThank you for sharing the parts of your solutions in the forum. \nI got your idea.
", "votes": 0, "replies": [] }, { "id": 269057, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-16T05:47:46.327000", "content": "@choker\nEdit:\nWhy did you post such a comment?\nIt's possible to sound uncomfortable.
", "votes": 0, "replies": [] }, { "id": 269122, "author_name": "JianchengGui", "author_url": "", "post_date": "2018-01-16T08:42:59.323000", "content": "It's much more meaningful then yours.
", "votes": 0, "replies": [] }, { "id": 269451, "author_name": "Josan", "author_url": "", "post_date": "2018-01-16T21:05:50.053000", "content": "I have to say this is my first \"real\" Kaggle competition. I don't know any technique like ensembling, feature selection, and etc. I do learning a lot from your shared model. Without you, the new beginners hard to get those techniques or tricks that commonly used in Kaggle competitions.
\n\nThose who feel it is unfair to people working hard on their own model are compassionate. If they know ensembling, even they don't want to learn the knowledge of the shared model, they should still gain benefits from it. For the detail of ensembling, please check https://www.kaggle.com/wiki/Ensembling
\n\nUpdated: \nPublic sharing is what the Rules supported: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/rules
", "votes": 1, "replies": [] } ] }, { "id": 268913, "author_name": "kirk", "author_url": "", "post_date": "2018-01-15T23:47:29.187000", "content": "Also mind the fact that 99% of the organizers who host a competition benefit more from discussion and ideas shared in the forums than from the actual top solution that might never end up in their production line. That's why some competitions as this one prevent us from sharing outside from the kaggle forums. Not to mention at all as other have said that here sharing and learning is what makes this community unique. Without that kaggle would be nothing.
", "votes": 6, "replies": [ { "id": 269060, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-16T05:51:11.610000", "content": "That's an interesting fact for me.
", "votes": 1, "replies": [] } ] }, { "id": 268857, "author_name": "Dieter", "author_url": "", "post_date": "2018-01-15T20:33:57.487000", "content": "i think it is pretty unfair towards people who try hard the last 3 months by themselves to reach top 10% and then someone shares recepies so everybody who has time in the last 2 weeks can easily reach 85%
", "votes": 4, "replies": [ { "id": 268862, "author_name": "Cut Onion", "author_url": "", "post_date": "2018-01-15T20:45:40.467000", "content": "Are you here for trying harder or learning better?
", "votes": 7, "replies": [] }, { "id": 269435, "author_name": "Piotr Czapla", "author_url": "", "post_date": "2018-01-16T20:46:49.067000", "content": "@kirk has fair point, even though it pains me that the exact code was shared. But then it make sense if ppl can build on that. Next time I would have to watch discussion more closely.
", "votes": 0, "replies": [] }, { "id": 269506, "author_name": "Cut Onion", "author_url": "", "post_date": "2018-01-16T23:55:04.580000", "content": "I've got another comment on email. And I have to say: almost nobody learns nothing after competition end. I don't know why is this but somehow competition setting makes you really want to learn and try thing to compete. It is clear that looking at previous competitions: there is no activity, nobody spends fulltime effort, nobody creates new techniques, nobody even repeats previous solutions for the most part. \nAnd ability to mess with leaderboard by disclosing good enough solutions is the part of fun being top performer here ^_^. Learn fast, do things, be first, don't care about what others doing.
", "votes": 6, "replies": [] }, { "id": 269513, "author_name": "inversion", "author_url": "", "post_date": "2018-01-17T00:13:42.360000", "content": "@Sergey - When I was competing, after every competition, I vowed to learn from the top solutions. But, in reality, I always just gave them a quick glance and then jumped into the next active competition. So, I totally agree with you!
", "votes": 5, "replies": [] }, { "id": 269542, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-17T01:13:08.513000", "content": "@Sergey Mushinskiy\nI understood it is important to share before the deadline by your post.\nThank you.
", "votes": 1, "replies": [] } ] }, { "id": 268742, "author_name": "Cut Onion", "author_url": "", "post_date": "2018-01-15T13:13:04.023000", "content": "Solution sharing may lead to more advanced solutions. I.e. if everybody shares different solutions reasonably far from competition deadline it make possible to develop new, original approaches on top of them.
", "votes": 4, "replies": [] }, { "id": 269852, "author_name": "hengck23", "author_url": "", "post_date": "2018-01-17T12:09:03.233000", "content": "@kakapo2
\n\nyou can also take a look at this, some people offer tips and solution for another kaggle competition. I learn a lot from them:
\n\nhttps://www.kaggle.com/c/statoil-iceberg-classifier-challenge/discussion/47062
\n\nin some case, even csv files are provided for experiments in ensemble:
\n\nhttps://www.kaggle.com/dongxu027/explore-stacking-lb-0-1463
", "votes": 1, "replies": [ { "id": 270245, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-18T00:21:39.803000", "content": "Thank you CherKeng, I'll read them.
", "votes": 0, "replies": [] } ] }, { "id": 269056, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-16T05:47:08.667000", "content": "Thank you all for answering my question.
\nOf course, I know it's a silly question, especially for experienced kagglers.
\nHowever, I believe it helps I or some new kagglers could understand that a competition in kaggle is not only a \"competition\".
Strategy Deception
", "votes": 2, "replies": [ { "id": 269062, "author_name": "yuusha_aaaaa", "author_url": "", "post_date": "2018-01-16T05:54:02.337000", "content": "Oh, do you mean some of posted solutions are aimed to mislead other competitiors?
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "268760": "There can only be few winners but everyone can move forward. My target is to move forward and not about winning. \n\nThink of winning in future, not now.", "268913": "Also mind the fact that 99% of the organizers who host a competition benefit more from discussion and ideas shared in the forums than from the actual top solution that might never end up in their production line. That's why some competitions as this one prevent us from sharing outside from the kaggle forums. Not to mention at all as other have said that here sharing and learning is what makes this community unique. Without that kaggle would be nothing.", "268857": "i think it is pretty unfair towards people who try hard the last 3 months by themselves to reach top 10% and then someone shares recepies so everybody who has time in the last 2 weeks can easily reach 85% ", "268742": "Solution sharing may lead to more advanced solutions. I.e. if everybody shares different solutions reasonably far from competition deadline it make possible to develop new, original approaches on top of them. ", "268720": "I'm new to kaggle competition and very curious some people are sharing the solutions competitive with the bronze ones in public LB **before the deadline.**\nAre there advantages to them or the competition?", "269852": "@kakapo2\n\nyou can also take a look at this, some people offer tips and solution for another kaggle competition. I learn a lot from them:\n\nhttps://www.kaggle.com/c/statoil-iceberg-classifier-challenge/discussion/47062\n\nin some case, even csv files are provided for experiments in ensemble:\n\n\nhttps://www.kaggle.com/dongxu027/explore-stacking-lb-0-1463\n", "269056": "Thank you all for answering my question. \nOf course, I know it's a silly question, especially for experienced kagglers. \nHowever, I believe it helps I or some new kagglers could understand that a competition in kaggle is not only a \"competition\".", "268728": "Strategy Deception" } }