{
  "id": 691279,
  "title": "Question about paper submission",
  "url": "/competitions/accident/discussion/691279",
  "author_name": "",
  "post_date": "2026-04-14T12:36:10.136059400Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello,</p>\n<p>Thank you for organizing the ACCIDENT @ CVPR Competition.</p>\n<p>I have a couple of questions regarding the expected scope of paper submissions:</p>\n<ol>\n<li>Regarding the reported performance, does the paper need to strictly reflect the final leaderboard submission (e.g., private test score or top-ranked result), or is it acceptable to base the paper on the most important method or idea that contributed to performance, even if it differs from the final submitted configuration?</li>\n<li>Is it acceptable to include improvements beyond the original competition submission (e.g., late submissions, a refined pipeline, or additional experiments), as long as they are clearly distinguished from the official submission?</li>\n</ol>\n<p>Thank you in advance.</p>",
  "messages": [
    {
      "id": "3441693",
      "postDate": "04/14/2026 12:36:10",
      "content": "<p>Hello,</p>\n<p>Thank you for organizing the ACCIDENT @ CVPR Competition.</p>\n<p>I have a couple of questions regarding the expected scope of paper submissions:</p>\n<ol>\n<li>Regarding the reported performance, does the paper need to strictly reflect the final leaderboard submission (e.g., private test score or top-ranked result), or is it acceptable to base the paper on the most important method or idea that contributed to performance, even if it differs from the final submitted configuration?</li>\n<li>Is it acceptable to include improvements beyond the original competition submission (e.g., late submissions, a refined pipeline, or additional experiments), as long as they are clearly distinguished from the official submission?</li>\n</ol>\n<p>Thank you in advance.</p>",
      "rawMarkdown": "Hello,\n\nThank you for organizing the ACCIDENT @ CVPR Competition.\n\nI have a couple of questions regarding the expected scope of paper submissions:\n\n1. Regarding the reported performance, does the paper need to strictly reflect the final leaderboard submission (e.g., private test score or top-ranked result), or is it acceptable to base the paper on the most important method or idea that contributed to performance, even if it differs from the final submitted configuration?\n2. Is it acceptable to include improvements beyond the original competition submission (e.g., late submissions, a refined pipeline, or additional experiments), as long as they are clearly distinguished from the official submission?\n\nThank you in advance.",
      "votes": null
    },
    {
      "id": "3441697",
      "postDate": "04/14/2026 12:44:54",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/choigu\" target=\"_blank\">@choigu</a> ,</p>\n<p>First of all, please always try to ask in the appropriate thread; do not make separate questions.\nIt is hard to navigate :)</p>\n<p>Anyway, the paper obviously can go beyond the final leaderboard. That is essentially why we spend a lot of time and effort on republishing everything and making it reusable.</p>\n<p>Also, only describe what makes sense to you.</p>\n<p>Best,\nLukas</p>",
      "rawMarkdown": "Hi @choigu ,\n\nFirst of all, please always try to ask in the appropriate thread; do not make separate questions.\nIt is hard to navigate :)\n\nAnyway, the paper obviously can go beyond the final leaderboard. That is essentially why we spend a lot of time and effort on republishing everything and making it reusable.\n\nAlso, only describe what makes sense to you.\n\nBest,\nLukas",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3441697,
      "author_name": "picekl",
      "author_url": "",
      "post_date": "04/14/2026 12:44:54",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/choigu\" target=\"_blank\">@choigu</a> ,</p>\n<p>First of all, please always try to ask in the appropriate thread; do not make separate questions.\nIt is hard to navigate :)</p>\n<p>Anyway, the paper obviously can go beyond the final leaderboard. That is essentially why we spend a lot of time and effort on republishing everything and making it reusable.</p>\n<p>Also, only describe what makes sense to you.</p>\n<p>Best,\nLukas</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3441693": "Hello,\n\nThank you for organizing the ACCIDENT @ CVPR Competition.\n\nI have a couple of questions regarding the expected scope of paper submissions:\n\n1. Regarding the reported performance, does the paper need to strictly reflect the final leaderboard submission (e.g., private test score or top-ranked result), or is it acceptable to base the paper on the most important method or idea that contributed to performance, even if it differs from the final submitted configuration?\n2. Is it acceptable to include improvements beyond the original competition submission (e.g., late submissions, a refined pipeline, or additional experiments), as long as they are clearly distinguished from the official submission?\n\nThank you in advance.",
    "3441697": "Hi @choigu ,\n\nFirst of all, please always try to ask in the appropriate thread; do not make separate questions.\nIt is hard to navigate :)\n\nAnyway, the paper obviously can go beyond the final leaderboard. That is essentially why we spend a lot of time and effort on republishing everything and making it reusable.\n\nAlso, only describe what makes sense to you.\n\nBest,\nLukas"
  },
  "source": "meta"
}