{
  "id": 68715,
  "title": "Anybody used AutoML?",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/68715",
  "author_name": "German Beyger",
  "post_date": "2018-10-16T11:49:02.412000",
  "votes": 5,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Did somebody tried to apply AutoML to this task? It will be intresting to listen about automl using experience.</p>",
  "messages": [
    {
      "id": 406403,
      "postDate": "2018-10-19T07:28:52.050Z",
      "content": "<p>i wonder if we just use  google autoML to make models automatically, is that a violation of the rule?</p>\n\n<p>It seems the autoML is using reinforcement learning for NAS (neural architectural search) which is better than human design network architecture. I am interested to know the performance of AutoML  ranking in this challenge </p>\n\n<p>(<a href=\"https://cloud.google.com/blog/products/gcp/automl-vision-in-action-from-ramen-to-branded-goods\">https://cloud.google.com/blog/products/gcp/automl-vision-in-action-from-ramen-to-branded-goods</a>)</p>",
      "rawMarkdown": "i wonder if we just use  google autoML to make models automatically, is that a violation of the rule?\n\nIt seems the autoML is using reinforcement learning for NAS (neural architectural search) which is better than human design network architecture. I am interested to know the performance of AutoML  ranking in this challenge \n\n(https://cloud.google.com/blog/products/gcp/automl-vision-in-action-from-ramen-to-branded-goods)",
      "votes": 5
    },
    {
      "id": 404795,
      "postDate": "2018-10-16T11:49:02.413Z",
      "content": "<p>Did somebody tried to apply AutoML to this task? It will be intresting to listen about automl using experience.</p>",
      "rawMarkdown": "Did somebody tried to apply AutoML to this task? It will be intresting to listen about automl using experience.",
      "votes": 5
    },
    {
      "id": 407757,
      "postDate": "2018-10-21T17:56:05.847Z",
      "content": "<p>+</p>",
      "rawMarkdown": "+"
    }
  ],
  "comments": [
    {
      "id": 406403,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "2018-10-19T07:28:52.050000",
      "content": "<p>i wonder if we just use  google autoML to make models automatically, is that a violation of the rule?</p>\n\n<p>It seems the autoML is using reinforcement learning for NAS (neural architectural search) which is better than human design network architecture. I am interested to know the performance of AutoML  ranking in this challenge </p>\n\n<p>(<a href=\"https://cloud.google.com/blog/products/gcp/automl-vision-in-action-from-ramen-to-branded-goods\">https://cloud.google.com/blog/products/gcp/automl-vision-in-action-from-ramen-to-branded-goods</a>)</p>",
      "votes": 5,
      "replies": []
    },
    {
      "id": 407757,
      "author_name": "Nazim Girach",
      "author_url": "",
      "post_date": "2018-10-21T17:56:05.847000",
      "content": "<p>+</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "406403": "i wonder if we just use  google autoML to make models automatically, is that a violation of the rule?\n\nIt seems the autoML is using reinforcement learning for NAS (neural architectural search) which is better than human design network architecture. I am interested to know the performance of AutoML  ranking in this challenge \n\n(https://cloud.google.com/blog/products/gcp/automl-vision-in-action-from-ramen-to-branded-goods)",
    "404795": "Did somebody tried to apply AutoML to this task? It will be intresting to listen about automl using experience.",
    "407757": "+"
  }
}