{
  "id": 100602,
  "title": "We should have to select only generative GAN for final submissions",
  "url": "/competitions/generative-dog-images/discussion/100602",
  "author_name": "",
  "post_date": "2019-07-19T13:58:36.091450700Z",
  "votes": null,
  "comment_count": 7,
  "views": 0,
  "content": "<p>I think that Kaggle team should ask us to select <strong>only generative GAN</strong>. \nIf we combine that with having to put our final submissions public and deactivate the automatic selection of the 2 best submissions (to avoid the problem with inactive kaggler) we could have a private LB approximately clean. </p>\n\n<p>I know that it would still be possible to cheat but I don't see the point or the interest. It's our duty to behave to encourage Kaggle to keep experimenting with this kind of new competition. </p>\n\n<p>And honestly, I prefer a competition with still a bit of interest even if a few people will use it to gain unfair medals than just accepting memorizing GAN and forgetting about real GAN. </p>\n\n<p>Anyway, it would be nice to hear more from Kaggle team.  </p>",
  "messages": [
    {
      "id": "579981",
      "postDate": "07/19/2019 13:58:36",
      "content": "<p>I think that Kaggle team should ask us to select <strong>only generative GAN</strong>. \nIf we combine that with having to put our final submissions public and deactivate the automatic selection of the 2 best submissions (to avoid the problem with inactive kaggler) we could have a private LB approximately clean. </p>\n\n<p>I know that it would still be possible to cheat but I don't see the point or the interest. It's our duty to behave to encourage Kaggle to keep experimenting with this kind of new competition. </p>\n\n<p>And honestly, I prefer a competition with still a bit of interest even if a few people will use it to gain unfair medals than just accepting memorizing GAN and forgetting about real GAN. </p>\n\n<p>Anyway, it would be nice to hear more from Kaggle team.  </p>",
      "rawMarkdown": "I think that Kaggle team should ask us to select **only generative GAN**. \nIf we combine that with having to put our final submissions public and deactivate the automatic selection of the 2 best submissions (to avoid the problem with inactive kaggler) we could have a private LB approximately clean. \n\nI know that it would still be possible to cheat but I don't see the point or the interest. It's our duty to behave to encourage Kaggle to keep experimenting with this kind of new competition. \n\nAnd honestly, I prefer a competition with still a bit of interest even if a few people will use it to gain unfair medals than just accepting memorizing GAN and forgetting about real GAN. \n\nAnyway, it would be nice to hear more from Kaggle team.",
      "votes": null
    },
    {
      "id": "579982",
      "postDate": "07/19/2019 13:59:13",
      "content": "<p>Sorry for coming back again and again with the same topic...</p>",
      "rawMarkdown": "Sorry for coming back again and again with the same topic...",
      "votes": null
    },
    {
      "id": "580476",
      "postDate": "07/20/2019 08:08:24",
      "content": "<p>why are people downvoting this ?</p>",
      "rawMarkdown": "why are people downvoting this ?",
      "votes": null
    },
    {
      "id": "580491",
      "postDate": "07/20/2019 08:33:54",
      "content": "<p>I think that this isn't feasible. Either make a clearly defined rule, or leave it as it is.\nSaying \"there is a way yo win, please don't use it, but if you use it, it is okay\" won't work.</p>",
      "rawMarkdown": "I think that this isn't feasible. Either make a clearly defined rule, or leave it as it is.\nSaying \"there is a way yo win, please don't use it, but if you use it, it is okay\" won't work.",
      "votes": null
    },
    {
      "id": "580498",
      "postDate": "07/20/2019 08:41:20",
      "content": "<p>I know there is no good solution here... But that's a shame that this competition end up wasted :(</p>",
      "rawMarkdown": "I know there is no good solution here... But that's a shame that this competition end up wasted :(",
      "votes": null
    },
    {
      "id": "580499",
      "postDate": "07/20/2019 08:41:31",
      "content": "<p>I'm wondering too ^^ </p>",
      "rawMarkdown": "I'm wondering too ^^",
      "votes": null
    },
    {
      "id": "580526",
      "postDate": "07/20/2019 09:52:35",
      "content": "<p>On the other hand it was a good experiment :) I hope it will help Kaggle team to create even better competitions in future.</p>",
      "rawMarkdown": "On the other hand it was a good experiment :) I hope it will help Kaggle team to create even better competitions in future.",
      "votes": null
    },
    {
      "id": "580531",
      "postDate": "07/20/2019 10:03:04",
      "content": "<p>fair enough :)</p>",
      "rawMarkdown": "fair enough :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 579982,
      "author_name": "francoisdubois",
      "author_url": "",
      "post_date": "07/19/2019 13:59:13",
      "content": "<p>Sorry for coming back again and again with the same topic...</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 580476,
      "author_name": "phoenix9032",
      "author_url": "",
      "post_date": "07/20/2019 08:08:24",
      "content": "<p>why are people downvoting this ?</p>",
      "votes": null,
      "replies": [
        {
          "id": 580499,
          "author_name": "francoisdubois",
          "author_url": "",
          "post_date": "07/20/2019 08:41:31",
          "content": "<p>I'm wondering too ^^ </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 580491,
      "author_name": "artgor",
      "author_url": "",
      "post_date": "07/20/2019 08:33:54",
      "content": "<p>I think that this isn't feasible. Either make a clearly defined rule, or leave it as it is.\nSaying \"there is a way yo win, please don't use it, but if you use it, it is okay\" won't work.</p>",
      "votes": null,
      "replies": [
        {
          "id": 580498,
          "author_name": "francoisdubois",
          "author_url": "",
          "post_date": "07/20/2019 08:41:20",
          "content": "<p>I know there is no good solution here... But that's a shame that this competition end up wasted :(</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 580526,
          "author_name": "artgor",
          "author_url": "",
          "post_date": "07/20/2019 09:52:35",
          "content": "<p>On the other hand it was a good experiment :) I hope it will help Kaggle team to create even better competitions in future.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 580531,
          "author_name": "francoisdubois",
          "author_url": "",
          "post_date": "07/20/2019 10:03:04",
          "content": "<p>fair enough :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "579981": "I think that Kaggle team should ask us to select **only generative GAN**. \nIf we combine that with having to put our final submissions public and deactivate the automatic selection of the 2 best submissions (to avoid the problem with inactive kaggler) we could have a private LB approximately clean. \n\nI know that it would still be possible to cheat but I don't see the point or the interest. It's our duty to behave to encourage Kaggle to keep experimenting with this kind of new competition. \n\nAnd honestly, I prefer a competition with still a bit of interest even if a few people will use it to gain unfair medals than just accepting memorizing GAN and forgetting about real GAN. \n\nAnyway, it would be nice to hear more from Kaggle team.",
    "579982": "Sorry for coming back again and again with the same topic...",
    "580476": "why are people downvoting this ?",
    "580491": "I think that this isn't feasible. Either make a clearly defined rule, or leave it as it is.\nSaying \"there is a way yo win, please don't use it, but if you use it, it is okay\" won't work.",
    "580498": "I know there is no good solution here... But that's a shame that this competition end up wasted :(",
    "580499": "I'm wondering too ^^",
    "580526": "On the other hand it was a good experiment :) I hope it will help Kaggle team to create even better competitions in future.",
    "580531": "fair enough :)"
  },
  "source": "meta"
}