{
  "id": 230613,
  "title": "Should Pseudo Label of public test dataset generated by pre-trained model be shared as external dataset?",
  "url": "/competitions/hubmap-kidney-segmentation/discussion/230613",
  "author_name": "",
  "post_date": "2021-04-04T20:29:51.464299500Z",
  "votes": 5,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?</p>",
  "messages": [
    {
      "id": "1262887",
      "postDate": "04/04/2021 20:29:51",
      "content": "<p>I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?</p>",
      "rawMarkdown": "I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?",
      "votes": null
    },
    {
      "id": "1262903",
      "postDate": "04/04/2021 21:11:00",
      "content": "<blockquote>\n  <p>I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?</p>\n</blockquote>\n<p>Yes,</p>\n<blockquote>\n  <p>Should Pseudo Label of public test dataset generated by pre-trained model be shared as external dataset?</p>\n</blockquote>\n<p>No.</p>",
      "rawMarkdown": "> I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?\n\nYes,\n\n> Should Pseudo Label of public test dataset generated by pre-trained model be shared as external dataset?\n\nNo.",
      "votes": null
    },
    {
      "id": "1285018",
      "postDate": "04/26/2021 13:56:32",
      "content": "<p>is this confirmed now? As I'm going through the threads I see conflicting answers.</p>",
      "rawMarkdown": "is this confirmed now? As I'm going through the threads I see conflicting answers.",
      "votes": null
    },
    {
      "id": "1285034",
      "postDate": "04/26/2021 14:08:23",
      "content": "<p>not sure if this is confirmed but pseudolabels are the work of your model and you would give away your work. </p>\n<p>on the other hand hand labels would lead to an unfair advantage and should probably shared level the playing field, i.e. the score should reflect deep learning skills and not glomeruli labeling skills. </p>",
      "rawMarkdown": "not sure if this is confirmed but pseudolabels are the work of your model and you would give away your work. \n\non the other hand hand labels would lead to an unfair advantage and should probably shared level the playing field, i.e. the score should reflect deep learning skills and not glomeruli labeling skills.",
      "votes": null
    },
    {
      "id": "1285933",
      "postDate": "04/27/2021 11:49:13",
      "content": "<p>See <a href=\"https://www.kaggle.com/addisonhoward\" target=\"_blank\">@addisonhoward</a> answer. <a href=\"https://www.kaggle.com/theoviel\" target=\"_blank\">@theoviel</a> is correct.</p>\n<p><a href=\"https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/227616#1256154\" target=\"_blank\">https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/227616#1256154</a></p>\n<blockquote>\n  <p>The data itself and notifying of its existence is considered publicly available external data. The hand-labeling performed by any individual is not required to be made public.</p>\n</blockquote>",
      "rawMarkdown": "See @addisonhoward answer. @theoviel is correct.\n\nhttps://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/227616#1256154\n\n> The data itself and notifying of its existence is considered publicly available external data. The hand-labeling performed by any individual is not required to be made public.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1262903,
      "author_name": "theoviel",
      "author_url": "",
      "post_date": "04/04/2021 21:11:00",
      "content": "<blockquote>\n  <p>I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?</p>\n</blockquote>\n<p>Yes,</p>\n<blockquote>\n  <p>Should Pseudo Label of public test dataset generated by pre-trained model be shared as external dataset?</p>\n</blockquote>\n<p>No.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1285018,
          "author_name": "andrasferenczi",
          "author_url": "",
          "post_date": "04/26/2021 13:56:32",
          "content": "<p>is this confirmed now? As I'm going through the threads I see conflicting answers.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1285034,
          "author_name": "rosuluc",
          "author_url": "",
          "post_date": "04/26/2021 14:08:23",
          "content": "<p>not sure if this is confirmed but pseudolabels are the work of your model and you would give away your work. </p>\n<p>on the other hand hand labels would lead to an unfair advantage and should probably shared level the playing field, i.e. the score should reflect deep learning skills and not glomeruli labeling skills. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1285933,
          "author_name": "mpware",
          "author_url": "",
          "post_date": "04/27/2021 11:49:13",
          "content": "<p>See <a href=\"https://www.kaggle.com/addisonhoward\" target=\"_blank\">@addisonhoward</a> answer. <a href=\"https://www.kaggle.com/theoviel\" target=\"_blank\">@theoviel</a> is correct.</p>\n<p><a href=\"https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/227616#1256154\" target=\"_blank\">https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/227616#1256154</a></p>\n<blockquote>\n  <p>The data itself and notifying of its existence is considered publicly available external data. The hand-labeling performed by any individual is not required to be made public.</p>\n</blockquote>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1262887": "I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?",
    "1262903": "> I used some pre-trained model to generate Pseudo Label on public test dataset, Are we allowed to do this?\n\nYes,\n\n> Should Pseudo Label of public test dataset generated by pre-trained model be shared as external dataset?\n\nNo.",
    "1285018": "is this confirmed now? As I'm going through the threads I see conflicting answers.",
    "1285034": "not sure if this is confirmed but pseudolabels are the work of your model and you would give away your work. \n\non the other hand hand labels would lead to an unfair advantage and should probably shared level the playing field, i.e. the score should reflect deep learning skills and not glomeruli labeling skills.",
    "1285933": "See @addisonhoward answer. @theoviel is correct.\n\nhttps://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/227616#1256154\n\n> The data itself and notifying of its existence is considered publicly available external data. The hand-labeling performed by any individual is not required to be made public."
  },
  "source": "meta"
}