{
  "id": 365440,
  "title": "How to break the 0.812 barrier？",
  "url": "/competitions/open-problems-multimodal/discussion/365440",
  "author_name": "E-Max AI",
  "post_date": "2022-11-11T08:38:51.230000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>For citeseq part:<br>\ngroup kfold by 3 donors (3 folds)<br>\nfour layers DNN with dropout<br>\nMean corr = 0.89447<br>\nOof corr   = 0.89447<br>\nMaybe it good enough</p>\n<p>but  I have no idea how to improve the   multiome part?</p>",
  "messages": [
    {
      "id": 2025464,
      "postDate": "2022-11-11T08:38:51.230Z",
      "content": "<p>For citeseq part:<br>\ngroup kfold by 3 donors (3 folds)<br>\nfour layers DNN with dropout<br>\nMean corr = 0.89447<br>\nOof corr   = 0.89447<br>\nMaybe it good enough</p>\n<p>but  I have no idea how to improve the   multiome part?</p>",
      "rawMarkdown": "For citeseq part:\ngroup kfold by 3 donors (3 folds)\nfour layers DNN with dropout\nMean corr = 0.89447\nOof corr   = 0.89447\nMaybe it good enough\n\nbut  I have no idea how to improve the   multiome part?",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2025464": "For citeseq part:\ngroup kfold by 3 donors (3 folds)\nfour layers DNN with dropout\nMean corr = 0.89447\nOof corr   = 0.89447\nMaybe it good enough\n\nbut  I have no idea how to improve the   multiome part?"
  }
}