{
  "id": 243052,
  "title": "Confusion on Datalabel annotation",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/243052",
  "author_name": "",
  "post_date": "2021-06-01T03:44:15.180444800Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi everyone,<br>\nI know this is a late discussion to ask about label annotation, but when i went through some of the publications where the abbreviations are marked as labels and in few it didnot.<br>\nFor example:</p>\n<ol>\n<li>Id: bf2742e0-b310-4df4-9c59-468ef394c3d9 <br>\n labels:  \"alzheimer s disease neuroimaging initiative adni\", \"adni\"</li>\n<li>Id: 80d54997-5704-4a50-8694-01ff432ebd14<br>\n labels: \"Agricultural Resource Management Survey\"<br>\n in the text given: The Agricultural Resource Management Survey (ARMS) is an annual survey <br>\n conducted</li>\n</ol>\n<p>Doesn't the second case should give, \"Agricultural Resource Management Survey\" and \"Agricultural Resource Management Survey ARMS\" ?</p>",
  "messages": [
    {
      "id": "1330672",
      "postDate": "06/01/2021 03:44:15",
      "content": "<p>Hi everyone,<br>\nI know this is a late discussion to ask about label annotation, but when i went through some of the publications where the abbreviations are marked as labels and in few it didnot.<br>\nFor example:</p>\n<ol>\n<li>Id: bf2742e0-b310-4df4-9c59-468ef394c3d9 <br>\n labels:  \"alzheimer s disease neuroimaging initiative adni\", \"adni\"</li>\n<li>Id: 80d54997-5704-4a50-8694-01ff432ebd14<br>\n labels: \"Agricultural Resource Management Survey\"<br>\n in the text given: The Agricultural Resource Management Survey (ARMS) is an annual survey <br>\n conducted</li>\n</ol>\n<p>Doesn't the second case should give, \"Agricultural Resource Management Survey\" and \"Agricultural Resource Management Survey ARMS\" ?</p>",
      "rawMarkdown": "Hi everyone,\nI know this is a late discussion to ask about label annotation, but when i went through some of the publications where the abbreviations are marked as labels and in few it didnot.\nFor example:\n1. Id: bf2742e0-b310-4df4-9c59-468ef394c3d9 \n     labels:  \"alzheimer s disease neuroimaging initiative adni\", \"adni\"\n2. Id: 80d54997-5704-4a50-8694-01ff432ebd14\n     labels: \"Agricultural Resource Management Survey\"\n     in the text given: The Agricultural Resource Management Survey (ARMS) is an annual survey \n     conducted\n\nDoesn't the second case should give, \"Agricultural Resource Management Survey\" and \"Agricultural Resource Management Survey ARMS\" ?",
      "votes": null
    },
    {
      "id": "1330692",
      "postDate": "06/01/2021 04:00:16",
      "content": "<p>The hosts have indicated that the training set is not exhaustively labeled, and since you only get the file id and the labels, you can't know which sentences they are actually referring to. If you look harder, you will find labels that do not exist in the file text and datasets in the text that are not in the labels. This is just part of the challenge.  If your model can pick out \"Agricultural Resource Management Survey\" or \"Agricultural Resource Management Survey ARMS\"  you will still get a true positive. Just don't put both for one sentence. </p>",
      "rawMarkdown": "The hosts have indicated that the training set is not exhaustively labeled, and since you only get the file id and the labels, you can't know which sentences they are actually referring to. If you look harder, you will find labels that do not exist in the file text and datasets in the text that are not in the labels. This is just part of the challenge.  If your model can pick out \"Agricultural Resource Management Survey\" or \"Agricultural Resource Management Survey ARMS\"  you will still get a true positive. Just don't put both for one sentence.",
      "votes": null
    },
    {
      "id": "1330734",
      "postDate": "06/01/2021 04:30:32",
      "content": "<p>thnks for the quick reply. understood.</p>",
      "rawMarkdown": "thnks for the quick reply. understood.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1330692,
      "author_name": "nbroad",
      "author_url": "",
      "post_date": "06/01/2021 04:00:16",
      "content": "<p>The hosts have indicated that the training set is not exhaustively labeled, and since you only get the file id and the labels, you can't know which sentences they are actually referring to. If you look harder, you will find labels that do not exist in the file text and datasets in the text that are not in the labels. This is just part of the challenge.  If your model can pick out \"Agricultural Resource Management Survey\" or \"Agricultural Resource Management Survey ARMS\"  you will still get a true positive. Just don't put both for one sentence. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1330734,
          "author_name": "narendra",
          "author_url": "",
          "post_date": "06/01/2021 04:30:32",
          "content": "<p>thnks for the quick reply. understood.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1330672": "Hi everyone,\nI know this is a late discussion to ask about label annotation, but when i went through some of the publications where the abbreviations are marked as labels and in few it didnot.\nFor example:\n1. Id: bf2742e0-b310-4df4-9c59-468ef394c3d9 \n     labels:  \"alzheimer s disease neuroimaging initiative adni\", \"adni\"\n2. Id: 80d54997-5704-4a50-8694-01ff432ebd14\n     labels: \"Agricultural Resource Management Survey\"\n     in the text given: The Agricultural Resource Management Survey (ARMS) is an annual survey \n     conducted\n\nDoesn't the second case should give, \"Agricultural Resource Management Survey\" and \"Agricultural Resource Management Survey ARMS\" ?",
    "1330692": "The hosts have indicated that the training set is not exhaustively labeled, and since you only get the file id and the labels, you can't know which sentences they are actually referring to. If you look harder, you will find labels that do not exist in the file text and datasets in the text that are not in the labels. This is just part of the challenge.  If your model can pick out \"Agricultural Resource Management Survey\" or \"Agricultural Resource Management Survey ARMS\"  you will still get a true positive. Just don't put both for one sentence.",
    "1330734": "thnks for the quick reply. understood."
  },
  "source": "meta"
}