{
  "id": 95222,
  "title": "Non-cheating kernels were selected by top public LB teams",
  "url": "/competitions/imet-2019-fgvc6/discussion/95222",
  "author_name": "",
  "post_date": "2019-06-10T22:22:08.861618800Z",
  "votes": 3,
  "comment_count": 7,
  "views": 0,
  "content": "<p>@backaggle mentioned in a comment</p>\n\n<blockquote>\n  <p>What happens if they do not select the submission/kernel that they are violated?. \n  If they drop their position in stage 2, it shows that they are scared the justice and select the non-cheating versions. \n  I cant wait to see the final result :)</p>\n</blockquote>\n\n<p>And it seems he's right. They didnt submit a solution using the external data but they could use it in a different way like many other have mentioned. \nHOW CAN JUSTICE BE ACHIEVED ? </p>",
  "messages": [
    {
      "id": "549629",
      "postDate": "06/10/2019 22:22:08",
      "content": "<p>@backaggle mentioned in a comment</p>\n\n<blockquote>\n  <p>What happens if they do not select the submission/kernel that they are violated?. \n  If they drop their position in stage 2, it shows that they are scared the justice and select the non-cheating versions. \n  I cant wait to see the final result :)</p>\n</blockquote>\n\n<p>And it seems he's right. They didnt submit a solution using the external data but they could use it in a different way like many other have mentioned. \nHOW CAN JUSTICE BE ACHIEVED ? </p>",
      "rawMarkdown": "backaggle mentioned in a comment\n&gt; What happens if they do not select the submission/kernel that they are violated?. \nIf they drop their position in stage 2, it shows that they are scared the justice and select the non-cheating versions. \nI cant wait to see the final result :)\n\nAnd it seems he's right. They didnt submit a solution using the external data but they could use it in a different way like many other have mentioned. \nHOW CAN JUSTICE BE ACHIEVED ?",
      "votes": null
    },
    {
      "id": "549685",
      "postDate": "06/10/2019 23:55:12",
      "content": "<p>We are all waiting for an official conclusion of what happened on the public LB!</p>",
      "rawMarkdown": "We are all waiting for an official conclusion of what happened on the public LB!",
      "votes": null
    },
    {
      "id": "549734",
      "postDate": "06/11/2019 01:35:07",
      "content": "<p><code>Non-cheating kernels were selected</code>\nHow do you know that? </p>",
      "rawMarkdown": "`Non-cheating kernels were selected`\nHow do you know that?",
      "votes": null
    },
    {
      "id": "549739",
      "postDate": "06/11/2019 01:43:27",
      "content": "<p>Private LB is correlated to the public one.\nthe socres of some teams dropped from 0.7+ to 0.67- how can explain that?</p>",
      "rawMarkdown": "Private LB is correlated to the public one.\nthe socres of some teams dropped from 0.7+ to 0.67- how can explain that?",
      "votes": null
    },
    {
      "id": "549750",
      "postDate": "06/11/2019 02:04:07",
      "content": "<p>I think if it is just normal overfitting to public LB, then why nobody explains how they overfit the public LB</p>",
      "rawMarkdown": "I think if it is just normal overfitting to public LB, then why nobody explains how they overfit the public LB",
      "votes": null
    },
    {
      "id": "549751",
      "postDate": "06/11/2019 02:05:13",
      "content": "<p>Many ways, e.g. blend 700+ with a standard 640, you'll get 670. Would you call it non-cheating?</p>",
      "rawMarkdown": "Many ways, e.g. blend 700+ with a standard 640, you'll get 670. Would you call it non-cheating?",
      "votes": null
    },
    {
      "id": "549756",
      "postDate": "06/11/2019 02:11:57",
      "content": "<p>why the only ones who overfitted the public LB are the ones who jumped in top 5 in the last 2 days?</p>",
      "rawMarkdown": "why the only ones who overfitted the public LB are the ones who jumped in top 5 in the last 2 days?",
      "votes": null
    },
    {
      "id": "551134",
      "postDate": "06/12/2019 11:33:01",
      "content": "<p>Probably duplicate accounts /multiple account holders they mean.. My inference came from <a href=\"https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings/discussion/18926#latest-108129\">https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings/discussion/18926#latest-108129</a> </p>",
      "rawMarkdown": "Probably duplicate accounts /multiple account holders they mean.. My inference came from https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings/discussion/18926#latest-108129",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 549685,
      "author_name": "strideradu",
      "author_url": "",
      "post_date": "06/10/2019 23:55:12",
      "content": "<p>We are all waiting for an official conclusion of what happened on the public LB!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 549734,
      "author_name": "artyomp",
      "author_url": "",
      "post_date": "06/11/2019 01:35:07",
      "content": "<p><code>Non-cheating kernels were selected</code>\nHow do you know that? </p>",
      "votes": null,
      "replies": [
        {
          "id": 549739,
          "author_name": "rinnqd",
          "author_url": "",
          "post_date": "06/11/2019 01:43:27",
          "content": "<p>Private LB is correlated to the public one.\nthe socres of some teams dropped from 0.7+ to 0.67- how can explain that?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 549750,
          "author_name": "strideradu",
          "author_url": "",
          "post_date": "06/11/2019 02:04:07",
          "content": "<p>I think if it is just normal overfitting to public LB, then why nobody explains how they overfit the public LB</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 549751,
          "author_name": "artyomp",
          "author_url": "",
          "post_date": "06/11/2019 02:05:13",
          "content": "<p>Many ways, e.g. blend 700+ with a standard 640, you'll get 670. Would you call it non-cheating?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 549756,
          "author_name": "rinnqd",
          "author_url": "",
          "post_date": "06/11/2019 02:11:57",
          "content": "<p>why the only ones who overfitted the public LB are the ones who jumped in top 5 in the last 2 days?</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 551134,
      "author_name": "naveenkumar1",
      "author_url": "",
      "post_date": "06/12/2019 11:33:01",
      "content": "<p>Probably duplicate accounts /multiple account holders they mean.. My inference came from <a href=\"https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings/discussion/18926#latest-108129\">https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings/discussion/18926#latest-108129</a> </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "549629": "backaggle mentioned in a comment\n&gt; What happens if they do not select the submission/kernel that they are violated?. \nIf they drop their position in stage 2, it shows that they are scared the justice and select the non-cheating versions. \nI cant wait to see the final result :)\n\nAnd it seems he's right. They didnt submit a solution using the external data but they could use it in a different way like many other have mentioned. \nHOW CAN JUSTICE BE ACHIEVED ?",
    "549685": "We are all waiting for an official conclusion of what happened on the public LB!",
    "549734": "`Non-cheating kernels were selected`\nHow do you know that?",
    "549739": "Private LB is correlated to the public one.\nthe socres of some teams dropped from 0.7+ to 0.67- how can explain that?",
    "549750": "I think if it is just normal overfitting to public LB, then why nobody explains how they overfit the public LB",
    "549751": "Many ways, e.g. blend 700+ with a standard 640, you'll get 670. Would you call it non-cheating?",
    "549756": "why the only ones who overfitted the public LB are the ones who jumped in top 5 in the last 2 days?",
    "551134": "Probably duplicate accounts /multiple account holders they mean.. My inference came from https://www.kaggle.com/c/airbnb-recruiting-new-user-bookings/discussion/18926#latest-108129"
  },
  "source": "meta"
}