{
  "id": 150183,
  "title": "Ultimate submissions Nash equilibrium problem",
  "url": "/competitions/flower-classification-with-tpus/discussion/150183",
  "author_name": "",
  "post_date": "2020-05-11T11:05:58.962326700Z",
  "votes": 5,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Because of the  kaggle official statement, additional data cannot be used in the final submission. Therefore, whether to use the additional data is a difficult choice.\n(1) If you use extra data for submission and others do not use the extra data for final submission, your ranking will rise, but if you get the top three, you will be disqualified because of official review of your data.\n(2) If you use additional data for submission and others also use the additional data for final submission, you do not get the top three, but your ranking is in the fourth or tenth place, you get a good ranking, and don't worry about disqualification, but if the official review is very careful, from top to bottom to check carefully, you will also be disqualified.\n(3) You did not use extra data for submission, and no one else, although you are not ranked high, but you have learned ML knowledge and technology, and do not have to worry about losing the competition qualification, and if the official can review some of the use of extra data Contestants, your ranking can also be improved.</p>\n\n<p>Therefore, the final submission is a Nash equilibrium issue. The most important point is whether the official can check the extra data very carefully. If the official can check each person's data very carefully (but it is difficult to do, if you use the extra data for pre-training and then use the model to train the game data, the official is difficult to check), we better choose not to use the The data model, if not, is really difficult to decide whether to use the additional data model as the final submission.</p>",
  "messages": [
    {
      "id": "842364",
      "postDate": "05/11/2020 11:05:58",
      "content": "<p>Because of the  kaggle official statement, additional data cannot be used in the final submission. Therefore, whether to use the additional data is a difficult choice.\n(1) If you use extra data for submission and others do not use the extra data for final submission, your ranking will rise, but if you get the top three, you will be disqualified because of official review of your data.\n(2) If you use additional data for submission and others also use the additional data for final submission, you do not get the top three, but your ranking is in the fourth or tenth place, you get a good ranking, and don't worry about disqualification, but if the official review is very careful, from top to bottom to check carefully, you will also be disqualified.\n(3) You did not use extra data for submission, and no one else, although you are not ranked high, but you have learned ML knowledge and technology, and do not have to worry about losing the competition qualification, and if the official can review some of the use of extra data Contestants, your ranking can also be improved.</p>\n\n<p>Therefore, the final submission is a Nash equilibrium issue. The most important point is whether the official can check the extra data very carefully. If the official can check each person's data very carefully (but it is difficult to do, if you use the extra data for pre-training and then use the model to train the game data, the official is difficult to check), we better choose not to use the The data model, if not, is really difficult to decide whether to use the additional data model as the final submission.</p>",
      "rawMarkdown": "Because of the  kaggle official statement, additional data cannot be used in the final submission. Therefore, whether to use the additional data is a difficult choice.\n(1) If you use extra data for submission and others do not use the extra data for final submission, your ranking will rise, but if you get the top three, you will be disqualified because of official review of your data.\n(2) If you use additional data for submission and others also use the additional data for final submission, you do not get the top three, but your ranking is in the fourth or tenth place, you get a good ranking, and don't worry about disqualification, but if the official review is very careful, from top to bottom to check carefully, you will also be disqualified.\n(3) You did not use extra data for submission, and no one else, although you are not ranked high, but you have learned ML knowledge and technology, and do not have to worry about losing the competition qualification, and if the official can review some of the use of extra data Contestants, your ranking can also be improved.\n\nTherefore, the final submission is a Nash equilibrium issue. The most important point is whether the official can check the extra data very carefully. If the official can check each person's data very carefully (but it is difficult to do, if you use the extra data for pre-training and then use the model to train the game data, the official is difficult to check), we better choose not to use the The data model, if not, is really difficult to decide whether to use the additional data model as the final submission.",
      "votes": null
    },
    {
      "id": "843262",
      "postDate": "05/11/2020 23:34:52",
      "content": "<p>Under the current conditions  my equilibrium is to strive off LB ambitions and to do my own experiments</p>",
      "rawMarkdown": "Under the current conditions  my equilibrium is to strive off LB ambitions and to do my own experiments",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 843262,
      "author_name": "romanweilguny",
      "author_url": "",
      "post_date": "05/11/2020 23:34:52",
      "content": "<p>Under the current conditions  my equilibrium is to strive off LB ambitions and to do my own experiments</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "842364": "Because of the  kaggle official statement, additional data cannot be used in the final submission. Therefore, whether to use the additional data is a difficult choice.\n(1) If you use extra data for submission and others do not use the extra data for final submission, your ranking will rise, but if you get the top three, you will be disqualified because of official review of your data.\n(2) If you use additional data for submission and others also use the additional data for final submission, you do not get the top three, but your ranking is in the fourth or tenth place, you get a good ranking, and don't worry about disqualification, but if the official review is very careful, from top to bottom to check carefully, you will also be disqualified.\n(3) You did not use extra data for submission, and no one else, although you are not ranked high, but you have learned ML knowledge and technology, and do not have to worry about losing the competition qualification, and if the official can review some of the use of extra data Contestants, your ranking can also be improved.\n\nTherefore, the final submission is a Nash equilibrium issue. The most important point is whether the official can check the extra data very carefully. If the official can check each person's data very carefully (but it is difficult to do, if you use the extra data for pre-training and then use the model to train the game data, the official is difficult to check), we better choose not to use the The data model, if not, is really difficult to decide whether to use the additional data model as the final submission.",
    "843262": "Under the current conditions  my equilibrium is to strive off LB ambitions and to do my own experiments"
  },
  "source": "meta"
}