{
  "id": 56847,
  "title": "Is hit_id showing time sequence of the hits?",
  "url": "/competitions/trackml-particle-identification/discussion/56847",
  "author_name": "",
  "post_date": "2018-05-15T18:03:14.573721700Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "",
  "messages": [
    {
      "id": "329102",
      "postDate": "05/15/2018 18:03:14",
      "content": "",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "329118",
      "postDate": "05/15/2018 18:55:12",
      "content": "<p>The hit_id numbering is completely independent from any hit timing. Within one event, all hits must be considered to occur simultaneously.</p>",
      "rawMarkdown": "The hit_id numbering is completely independent from any hit timing. Within one event, all hits must be considered to occur simultaneously.",
      "votes": null
    },
    {
      "id": "329286",
      "postDate": "05/16/2018 06:36:45",
      "content": "<p>@Moritz Kiehn - if all the hits are considered to occur simultaneously how is a cell able to make a distinction between multiple hits? </p>\n\n<p>Take for example event000001000 that for volume_id=13, layer_id=4, module_id=451, ch0=92 and 93, ch1 =42 has 5 hits? (Considering this volume's modules have strip cells and only record a value of 1 representing that a particle generated a hit.) There are 802 cells that recorder more than 1 hit for that event. </p>",
      "rawMarkdown": "Moritz Kiehn - if all the hits are considered to occur simultaneously how is a cell able to make a distinction between multiple hits? \n\nTake for example event000001000 that for volume_id=13, layer_id=4, module_id=451, ch0=92 and 93, ch1 =42 has 5 hits? (Considering this volume's modules have strip cells and only record a value of 1 representing that a particle generated a hit.) There are 802 cells that recorder more than 1 hit for that event.",
      "votes": null
    },
    {
      "id": "329327",
      "postDate": "05/16/2018 08:36:24",
      "content": "<p>This is a simplification we did for the dataset: in reality, these cells would be merged (although we have a cluster splitter in place). However, to allow unique labeling, i.e. a unique association from each hit to exactly one track.</p>\n\n<p>This was necessary in order to keep the scoring function simple enough that it can be executed in a reasonable time. As it is, almost every problem we are tackling in this environment is somehow CPU bound/limited.</p>",
      "rawMarkdown": "This is a simplification we did for the dataset: in reality, these cells would be merged (although we have a cluster splitter in place). However, to allow unique labeling, i.e. a unique association from each hit to exactly one track.\n\nThis was necessary in order to keep the scoring function simple enough that it can be executed in a reasonable time. As it is, almost every problem we are tackling in this environment is somehow CPU bound/limited.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 329118,
      "author_name": "msmk00",
      "author_url": "",
      "post_date": "05/15/2018 18:55:12",
      "content": "<p>The hit_id numbering is completely independent from any hit timing. Within one event, all hits must be considered to occur simultaneously.</p>",
      "votes": null,
      "replies": [
        {
          "id": 329286,
          "author_name": "profetul",
          "author_url": "",
          "post_date": "05/16/2018 06:36:45",
          "content": "<p>@Moritz Kiehn - if all the hits are considered to occur simultaneously how is a cell able to make a distinction between multiple hits? </p>\n\n<p>Take for example event000001000 that for volume_id=13, layer_id=4, module_id=451, ch0=92 and 93, ch1 =42 has 5 hits? (Considering this volume's modules have strip cells and only record a value of 1 representing that a particle generated a hit.) There are 802 cells that recorder more than 1 hit for that event. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 329327,
          "author_name": "asalzburger",
          "author_url": "",
          "post_date": "05/16/2018 08:36:24",
          "content": "<p>This is a simplification we did for the dataset: in reality, these cells would be merged (although we have a cluster splitter in place). However, to allow unique labeling, i.e. a unique association from each hit to exactly one track.</p>\n\n<p>This was necessary in order to keep the scoring function simple enough that it can be executed in a reasonable time. As it is, almost every problem we are tackling in this environment is somehow CPU bound/limited.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "329102": "",
    "329118": "The hit_id numbering is completely independent from any hit timing. Within one event, all hits must be considered to occur simultaneously.",
    "329286": "Moritz Kiehn - if all the hits are considered to occur simultaneously how is a cell able to make a distinction between multiple hits? \n\nTake for example event000001000 that for volume_id=13, layer_id=4, module_id=451, ch0=92 and 93, ch1 =42 has 5 hits? (Considering this volume's modules have strip cells and only record a value of 1 representing that a particle generated a hit.) There are 802 cells that recorder more than 1 hit for that event.",
    "329327": "This is a simplification we did for the dataset: in reality, these cells would be merged (although we have a cluster splitter in place). However, to allow unique labeling, i.e. a unique association from each hit to exactly one track.\n\nThis was necessary in order to keep the scoring function simple enough that it can be executed in a reasonable time. As it is, almost every problem we are tackling in this environment is somehow CPU bound/limited."
  },
  "source": "meta"
}