{
  "id": 963,
  "title": "And the winners are...",
  "url": "/competitions/SemiSupervisedFeatureLearning/discussion/963",
  "author_name": "D. Sculley",
  "post_date": "2011-10-19T11:26:42.863000",
  "votes": 0,
  "comment_count": 0,
  "views": 2650,
  "content": "<p>Congratulations to Benjamin Hamner and William Cukierski for winning this competition! &nbsp;We've verfied the results and are pleased to award the prize.</p>\r\n<p>Benjamin and William have posted some details of their winning method on the &quot;Contest methods&quot; thread in the forum. &nbsp;In particular, it's surprising and interesting that their method of using multiple supervised random forests beat out a range of other methods\r\n including deep learning methods, label propagation methods, sparse coding methods, and unsupervised and semi-supervised feature learning methods.</p>\r\n<p>We're finishing up the full writeup of the competition and should be able to post a preliminary version of this in a few days.</p>\r\n<p>Thanks, again, to everyone who participated and submitted results!</p>",
  "messages": [
    {
      "id": 6151,
      "postDate": "2011-10-19T11:26:42.863Z",
      "content": "<p>Congratulations to Benjamin Hamner and William Cukierski for winning this competition! &nbsp;We've verfied the results and are pleased to award the prize.</p>\r\n<p>Benjamin and William have posted some details of their winning method on the &quot;Contest methods&quot; thread in the forum. &nbsp;In particular, it's surprising and interesting that their method of using multiple supervised random forests beat out a range of other methods\r\n including deep learning methods, label propagation methods, sparse coding methods, and unsupervised and semi-supervised feature learning methods.</p>\r\n<p>We're finishing up the full writeup of the competition and should be able to post a preliminary version of this in a few days.</p>\r\n<p>Thanks, again, to everyone who participated and submitted results!</p>"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "6151": ""
  }
}