{
  "id": 60203,
  "title": "what is a good unique descriptor for each track?",
  "url": "/competitions/trackml-particle-identification/discussion/60203",
  "author_name": "",
  "post_date": "2018-07-01T17:27:59.609947800Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I am trying to train a deep cnn to output unique descriptor for each track.\nI need target values (ground truth values) for the deep cnn to learn.</p>\n\n<p>Is it correct to say that in a single event:</p>\n\n<ul>\n<li><p>all tracks will have completely different track parameters (helix parameters r0, theat0, gamma)</p></li>\n<li><p>all tracks will have completely different (initial velocity, initial momentum)</p></li>\n<li><p>all tracks will have completely different (initial location, final location)</p></li>\n</ul>",
  "messages": [
    {
      "id": "351239",
      "postDate": "07/01/2018 17:27:59",
      "content": "<p>I am trying to train a deep cnn to output unique descriptor for each track.\nI need target values (ground truth values) for the deep cnn to learn.</p>\n\n<p>Is it correct to say that in a single event:</p>\n\n<ul>\n<li><p>all tracks will have completely different track parameters (helix parameters r0, theat0, gamma)</p></li>\n<li><p>all tracks will have completely different (initial velocity, initial momentum)</p></li>\n<li><p>all tracks will have completely different (initial location, final location)</p></li>\n</ul>",
      "rawMarkdown": "I am trying to train a deep cnn to output unique descriptor for each track.\nI need target values (ground truth values) for the deep cnn to learn.\n\nIs it correct to say that in a single event:\n\n  - all tracks will have completely different track parameters (helix parameters r0, theat0, gamma)\n\n  - all tracks will have completely different (initial velocity, initial momentum)\n\n  - all tracks will have completely different (initial location, final location)",
      "votes": null
    },
    {
      "id": "351250",
      "postDate": "07/01/2018 17:50:49",
      "content": "<p>strictly speaking NO: in particular  an \"initial location\" should be the very same  for several tracks\n( try <code>cat event000001003-particles.csv | cut -d, -f2- | sort -n | less</code>)\nbesides the common origin the other parameters are as different as in any random generator</p>",
      "rawMarkdown": "strictly speaking NO: in particular  an \"initial location\" should be the very same  for several tracks\n( try ```cat event000001003-particles.csv | cut -d, -f2- | sort -n | less```)\nbesides the common origin the other parameters are as different as in any random generator",
      "votes": null
    },
    {
      "id": "351479",
      "postDate": "07/02/2018 09:25:33",
      "content": "<p>I concur, particles are created from a collision, and a collision can generate many particles.  It means that for each origin there are a number of tracks.  The rest of the features are probably unique to each track.</p>",
      "rawMarkdown": "I concur, particles are created from a collision, and a collision can generate many particles.  It means that for each origin there are a number of tracks.  The rest of the features are probably unique to each track.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 351250,
      "author_name": "vininn",
      "author_url": "",
      "post_date": "07/01/2018 17:50:49",
      "content": "<p>strictly speaking NO: in particular  an \"initial location\" should be the very same  for several tracks\n( try <code>cat event000001003-particles.csv | cut -d, -f2- | sort -n | less</code>)\nbesides the common origin the other parameters are as different as in any random generator</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 351479,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "07/02/2018 09:25:33",
      "content": "<p>I concur, particles are created from a collision, and a collision can generate many particles.  It means that for each origin there are a number of tracks.  The rest of the features are probably unique to each track.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "351239": "I am trying to train a deep cnn to output unique descriptor for each track.\nI need target values (ground truth values) for the deep cnn to learn.\n\nIs it correct to say that in a single event:\n\n  - all tracks will have completely different track parameters (helix parameters r0, theat0, gamma)\n\n  - all tracks will have completely different (initial velocity, initial momentum)\n\n  - all tracks will have completely different (initial location, final location)",
    "351250": "strictly speaking NO: in particular  an \"initial location\" should be the very same  for several tracks\n( try ```cat event000001003-particles.csv | cut -d, -f2- | sort -n | less```)\nbesides the common origin the other parameters are as different as in any random generator",
    "351479": "I concur, particles are created from a collision, and a collision can generate many particles.  It means that for each origin there are a number of tracks.  The rest of the features are probably unique to each track."
  },
  "source": "meta"
}