{
  "id": 57227,
  "title": "DCASE2018 Challenge: Acoustic scene classification tasks and parallel Kaggle InClass competitions",
  "url": "/competitions/acoustic-scene-2018/discussion/57227",
  "author_name": "Toni Heittola",
  "post_date": "2018-05-21T08:11:50.030000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi all,</p>\n\n<p>This year's edition of DCASE Challenge contains again task for Acoustic scene classification in very similar setup to this competition. We have just opened parallel Kaggle InClass competitions for the subtasks under the challenge, and you are invited to join the competitions. The aim of these Kaggle competitions is to allow challenge participants test their system with unseen data, and to compete against each other in public leaderboard during the development phase of the challenge. These parallel competitions are open to everybody and submitting to them does not require participation to actual DCASE challenge. So if you are interested on tackling the acoustic scene classification problem you are welcome to join us!</p>\n\n<p>There are three subtasks approaching the scene classification problem from a slightly different direction: A) scene classification with high-quality binaural (stereo) recordings, B) scene classification when your material contains recordings from various devices (mobile device), and C) scene classification with use of external material. Setup in the subtask A is very similar to this Kaggle competition (\"TUT Acoustic Scene Classification\"), only difference being that the material is released as audio, not as pre-calculated audio features, and instead of 15 scene classes, there are now 10 scene classes.</p>\n\n<ul>\n<li>Task1A Leaderboard &gt; <a href=\"https://www.kaggle.com/c/dcase2018-task1a-leaderboard\">https://www.kaggle.com/c/dcase2018-task1a-leaderboard</a></li>\n<li>Task1B Leaderboard &gt; <a href=\"https://www.kaggle.com/c/dcase2018-task1b-leaderboard\">https://www.kaggle.com/c/dcase2018-task1b-leaderboard</a></li>\n<li>Task1C Leaderboard &gt; <a href=\"https://www.kaggle.com/c/dcase2018-task1c-leaderboard\">https://www.kaggle.com/c/dcase2018-task1c-leaderboard</a></li>\n</ul>\n\n<p>Best regards,</p>\n\n<p>Toni Heittola</p>\n\n<p>DCASE2018 Task 1 coordinator </p>",
  "messages": [
    {
      "id": 331434,
      "postDate": "2018-05-21T08:11:50.030Z",
      "content": "<p>Hi all,</p>\n\n<p>This year's edition of DCASE Challenge contains again task for Acoustic scene classification in very similar setup to this competition. We have just opened parallel Kaggle InClass competitions for the subtasks under the challenge, and you are invited to join the competitions. The aim of these Kaggle competitions is to allow challenge participants test their system with unseen data, and to compete against each other in public leaderboard during the development phase of the challenge. These parallel competitions are open to everybody and submitting to them does not require participation to actual DCASE challenge. So if you are interested on tackling the acoustic scene classification problem you are welcome to join us!</p>\n\n<p>There are three subtasks approaching the scene classification problem from a slightly different direction: A) scene classification with high-quality binaural (stereo) recordings, B) scene classification when your material contains recordings from various devices (mobile device), and C) scene classification with use of external material. Setup in the subtask A is very similar to this Kaggle competition (\"TUT Acoustic Scene Classification\"), only difference being that the material is released as audio, not as pre-calculated audio features, and instead of 15 scene classes, there are now 10 scene classes.</p>\n\n<ul>\n<li>Task1A Leaderboard &gt; <a href=\"https://www.kaggle.com/c/dcase2018-task1a-leaderboard\">https://www.kaggle.com/c/dcase2018-task1a-leaderboard</a></li>\n<li>Task1B Leaderboard &gt; <a href=\"https://www.kaggle.com/c/dcase2018-task1b-leaderboard\">https://www.kaggle.com/c/dcase2018-task1b-leaderboard</a></li>\n<li>Task1C Leaderboard &gt; <a href=\"https://www.kaggle.com/c/dcase2018-task1c-leaderboard\">https://www.kaggle.com/c/dcase2018-task1c-leaderboard</a></li>\n</ul>\n\n<p>Best regards,</p>\n\n<p>Toni Heittola</p>\n\n<p>DCASE2018 Task 1 coordinator </p>",
      "rawMarkdown": "Hi all,\n\nThis year's edition of DCASE Challenge contains again task for Acoustic scene classification in very similar setup to this competition. We have just opened parallel Kaggle InClass competitions for the subtasks under the challenge, and you are invited to join the competitions. The aim of these Kaggle competitions is to allow challenge participants test their system with unseen data, and to compete against each other in public leaderboard during the development phase of the challenge. These parallel competitions are open to everybody and submitting to them does not require participation to actual DCASE challenge. So if you are interested on tackling the acoustic scene classification problem you are welcome to join us!\n\nThere are three subtasks approaching the scene classification problem from a slightly different direction: A) scene classification with high-quality binaural (stereo) recordings, B) scene classification when your material contains recordings from various devices (mobile device), and C) scene classification with use of external material. Setup in the subtask A is very similar to this Kaggle competition (\"TUT Acoustic Scene Classification\"), only difference being that the material is released as audio, not as pre-calculated audio features, and instead of 15 scene classes, there are now 10 scene classes.\n\n- Task1A Leaderboard &gt; https://www.kaggle.com/c/dcase2018-task1a-leaderboard\n- Task1B Leaderboard &gt; https://www.kaggle.com/c/dcase2018-task1b-leaderboard\n- Task1C Leaderboard &gt; https://www.kaggle.com/c/dcase2018-task1c-leaderboard\n\nBest regards,\n\nToni Heittola\n\nDCASE2018 Task 1 coordinator "
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "331434": "Hi all,\n\nThis year's edition of DCASE Challenge contains again task for Acoustic scene classification in very similar setup to this competition. We have just opened parallel Kaggle InClass competitions for the subtasks under the challenge, and you are invited to join the competitions. The aim of these Kaggle competitions is to allow challenge participants test their system with unseen data, and to compete against each other in public leaderboard during the development phase of the challenge. These parallel competitions are open to everybody and submitting to them does not require participation to actual DCASE challenge. So if you are interested on tackling the acoustic scene classification problem you are welcome to join us!\n\nThere are three subtasks approaching the scene classification problem from a slightly different direction: A) scene classification with high-quality binaural (stereo) recordings, B) scene classification when your material contains recordings from various devices (mobile device), and C) scene classification with use of external material. Setup in the subtask A is very similar to this Kaggle competition (\"TUT Acoustic Scene Classification\"), only difference being that the material is released as audio, not as pre-calculated audio features, and instead of 15 scene classes, there are now 10 scene classes.\n\n- Task1A Leaderboard &gt; https://www.kaggle.com/c/dcase2018-task1a-leaderboard\n- Task1B Leaderboard &gt; https://www.kaggle.com/c/dcase2018-task1b-leaderboard\n- Task1C Leaderboard &gt; https://www.kaggle.com/c/dcase2018-task1c-leaderboard\n\nBest regards,\n\nToni Heittola\n\nDCASE2018 Task 1 coordinator "
  }
}