{
  "id": 388329,
  "title": "What part of this is the hard part?",
  "url": "/competitions/icecube-neutrinos-in-deep-ice/discussion/388329",
  "author_name": "",
  "post_date": "2023-02-16T22:15:19.999025600Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>What are the primary reasons that this task is computationally intensive? So far, I have not downloaded the data, but I'm trying to get a better idea of what problem needs to be solved.</p>",
  "messages": [
    {
      "id": "2147807",
      "postDate": "02/16/2023 22:15:20",
      "content": "<p>What are the primary reasons that this task is computationally intensive? So far, I have not downloaded the data, but I'm trying to get a better idea of what problem needs to be solved.</p>",
      "rawMarkdown": "What are the primary reasons that this task is computationally intensive? So far, I have not downloaded the data, but I'm trying to get a better idea of what problem needs to be solved.",
      "votes": null
    },
    {
      "id": "2147945",
      "postDate": "02/17/2023 03:19:03",
      "content": "<p>The data set is large - but for the most part 99% of what you will do can be performed on 1 of the 661 training batches.  </p>\n<p>Problem 1 - find a math solution to determine direction.  Several shared notebooks that show this.  A decent part of the test data can arrive at a pretty accurate value using math only.   But decent won't win any prizes ??  Lots of noise added to the data.</p>\n<p>Problem 2 - find a AI tool solution to determine direction.  The top scoring shared notebooks show this.  Once you have a good model training it becomes computationally intensive.</p>\n<p>Problem 3 - both 1 and 2 have to fight off the noise.  IMO the problem to be solved is learning to identify and ignore the noise and use math and/or AI to determine direction.</p>\n<p>The Icecube has a large number of very smart people who have been working the issue for a decade or more.  </p>",
      "rawMarkdown": "The data set is large - but for the most part 99% of what you will do can be performed on 1 of the 661 training batches.  \n\nProblem 1 - find a math solution to determine direction.  Several shared notebooks that show this.  A decent part of the test data can arrive at a pretty accurate value using math only.   But decent won't win any prizes ??  Lots of noise added to the data.\n\nProblem 2 - find a AI tool solution to determine direction.  The top scoring shared notebooks show this.  Once you have a good model training it becomes computationally intensive.\n\nProblem 3 - both 1 and 2 have to fight off the noise.  IMO the problem to be solved is learning to identify and ignore the noise and use math and/or AI to determine direction.\n\nThe Icecube has a large number of very smart people who have been working the issue for a decade or more.",
      "votes": null
    },
    {
      "id": "2150101",
      "postDate": "02/18/2023 23:56:41",
      "content": "<p>I recommend watching the webinar, it gives a great overview of the problem :)</p>\n<p><a href=\"https://www.youtube.com/watch?v=_G4Nx20BVPg\" target=\"_blank\">https://www.youtube.com/watch?v=_G4Nx20BVPg</a></p>",
      "rawMarkdown": "I recommend watching the webinar, it gives a great overview of the problem :)\n\nhttps://www.youtube.com/watch?v=_G4Nx20BVPg",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2147945,
      "author_name": "pcjimmmy",
      "author_url": "",
      "post_date": "02/17/2023 03:19:03",
      "content": "<p>The data set is large - but for the most part 99% of what you will do can be performed on 1 of the 661 training batches.  </p>\n<p>Problem 1 - find a math solution to determine direction.  Several shared notebooks that show this.  A decent part of the test data can arrive at a pretty accurate value using math only.   But decent won't win any prizes ??  Lots of noise added to the data.</p>\n<p>Problem 2 - find a AI tool solution to determine direction.  The top scoring shared notebooks show this.  Once you have a good model training it becomes computationally intensive.</p>\n<p>Problem 3 - both 1 and 2 have to fight off the noise.  IMO the problem to be solved is learning to identify and ignore the noise and use math and/or AI to determine direction.</p>\n<p>The Icecube has a large number of very smart people who have been working the issue for a decade or more.  </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2150101,
      "author_name": "araraonline",
      "author_url": "",
      "post_date": "02/18/2023 23:56:41",
      "content": "<p>I recommend watching the webinar, it gives a great overview of the problem :)</p>\n<p><a href=\"https://www.youtube.com/watch?v=_G4Nx20BVPg\" target=\"_blank\">https://www.youtube.com/watch?v=_G4Nx20BVPg</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2147807": "What are the primary reasons that this task is computationally intensive? So far, I have not downloaded the data, but I'm trying to get a better idea of what problem needs to be solved.",
    "2147945": "The data set is large - but for the most part 99% of what you will do can be performed on 1 of the 661 training batches.  \n\nProblem 1 - find a math solution to determine direction.  Several shared notebooks that show this.  A decent part of the test data can arrive at a pretty accurate value using math only.   But decent won't win any prizes ??  Lots of noise added to the data.\n\nProblem 2 - find a AI tool solution to determine direction.  The top scoring shared notebooks show this.  Once you have a good model training it becomes computationally intensive.\n\nProblem 3 - both 1 and 2 have to fight off the noise.  IMO the problem to be solved is learning to identify and ignore the noise and use math and/or AI to determine direction.\n\nThe Icecube has a large number of very smart people who have been working the issue for a decade or more.",
    "2150101": "I recommend watching the webinar, it gives a great overview of the problem :)\n\nhttps://www.youtube.com/watch?v=_G4Nx20BVPg"
  },
  "source": "meta"
}