{
  "id": 457575,
  "title": "Flipping the sequence as a form of data augmentation?",
  "url": "/competitions/stanford-ribonanza-rna-folding/discussion/457575",
  "author_name": "",
  "post_date": "2023-11-25T13:54:37.938388100Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hello, I am wondering if anyone has tried to flip the RNA sequence as a form of data augmentation ?</p>\n<p>For example, you could take sequence \"GGGAACGA\" and flip it, with its corresponding reactivities, to \"AGCAAGGG\" to effectively double the number of training samples.</p>\n<p>This type of augmentation is very common in vision tasks, but I'm not sure if it would be appropriate here. Might it destroy some important information?</p>\n<p>Curious to hear different thoughts, and how other people have approached data augmentation for this competition.</p>",
  "messages": [
    {
      "id": "2537805",
      "postDate": "11/25/2023 13:54:37",
      "content": "<p>Hello, I am wondering if anyone has tried to flip the RNA sequence as a form of data augmentation ?</p>\n<p>For example, you could take sequence \"GGGAACGA\" and flip it, with its corresponding reactivities, to \"AGCAAGGG\" to effectively double the number of training samples.</p>\n<p>This type of augmentation is very common in vision tasks, but I'm not sure if it would be appropriate here. Might it destroy some important information?</p>\n<p>Curious to hear different thoughts, and how other people have approached data augmentation for this competition.</p>",
      "rawMarkdown": "Hello, I am wondering if anyone has tried to flip the RNA sequence as a form of data augmentation ?\n\nFor example, you could take sequence \"GGGAACGA\" and flip it, with its corresponding reactivities, to \"AGCAAGGG\" to effectively double the number of training samples.\n\nThis type of augmentation is very common in vision tasks, but I'm not sure if it would be appropriate here. Might it destroy some important information?\n\nCurious to hear different thoughts, and how other people have approached data augmentation for this competition.",
      "votes": null
    },
    {
      "id": "2537860",
      "postDate": "11/25/2023 14:53:48",
      "content": "<p>You can see the following thread <a href=\"https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/454999\" target=\"_blank\">https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/454999</a></p>",
      "rawMarkdown": "You can see the following thread https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/454999",
      "votes": null
    },
    {
      "id": "2538557",
      "postDate": "11/26/2023 09:04:46",
      "content": "<p>Since the RNA chain has an innate directionality, such a flipping doesn't seem to make scientific sense.</p>",
      "rawMarkdown": "Since the RNA chain has an innate directionality, such a flipping doesn't seem to make scientific sense.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2537860,
      "author_name": "dmitrypenzar1996",
      "author_url": "",
      "post_date": "11/25/2023 14:53:48",
      "content": "<p>You can see the following thread <a href=\"https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/454999\" target=\"_blank\">https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/454999</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2538557,
      "author_name": "jbomitchell",
      "author_url": "",
      "post_date": "11/26/2023 09:04:46",
      "content": "<p>Since the RNA chain has an innate directionality, such a flipping doesn't seem to make scientific sense.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2537805": "Hello, I am wondering if anyone has tried to flip the RNA sequence as a form of data augmentation ?\n\nFor example, you could take sequence \"GGGAACGA\" and flip it, with its corresponding reactivities, to \"AGCAAGGG\" to effectively double the number of training samples.\n\nThis type of augmentation is very common in vision tasks, but I'm not sure if it would be appropriate here. Might it destroy some important information?\n\nCurious to hear different thoughts, and how other people have approached data augmentation for this competition.",
    "2537860": "You can see the following thread https://www.kaggle.com/competitions/stanford-ribonanza-rna-folding/discussion/454999",
    "2538557": "Since the RNA chain has an innate directionality, such a flipping doesn't seem to make scientific sense."
  },
  "source": "meta"
}