{
  "id": 179063,
  "title": "Is audio overlapping essentially \"Mix-up\" for audio?",
  "url": "/competitions/birdsong-recognition/discussion/179063",
  "author_name": "",
  "post_date": "2020-09-01T09:11:44.354839300Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I had great result with \"Mix-up\" for supervised computer vision problems. I wonder if \"Mix-up\" can do the same thing for audio.</p>\n<pre><code>mixed_up = a*wav_1 + b*wav_2\nnew_labels = a*label_1 + b*label_2\n</code></pre>",
  "messages": [
    {
      "id": "993961",
      "postDate": "09/01/2020 09:11:44",
      "content": "<p>I had great result with \"Mix-up\" for supervised computer vision problems. I wonder if \"Mix-up\" can do the same thing for audio.</p>\n<pre><code>mixed_up = a*wav_1 + b*wav_2\nnew_labels = a*label_1 + b*label_2\n</code></pre>",
      "rawMarkdown": "I had great result with \"Mix-up\" for supervised computer vision problems. I wonder if \"Mix-up\" can do the same thing for audio.\n```\nmixed_up = a*wav_1 + b*wav_2\nnew_labels = a*label_1 + b*label_2\n```",
      "votes": null
    },
    {
      "id": "995333",
      "postDate": "09/02/2020 11:12:00",
      "content": "<p>Well, I think it's worth a try. We do it all the time and it definitely diversifies the audio data. You might wanna think about a more elaborate way of doing this (e.g., weighted sum) but mix-up in general should work just fine. </p>",
      "rawMarkdown": "Well, I think it's worth a try. We do it all the time and it definitely diversifies the audio data. You might wanna think about a more elaborate way of doing this (e.g., weighted sum) but mix-up in general should work just fine.",
      "votes": null
    },
    {
      "id": "995393",
      "postDate": "09/02/2020 12:04:16",
      "content": "<p>I definately will give it a shot!</p>",
      "rawMarkdown": "I definately will give it a shot!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 995333,
      "author_name": "stefankahl",
      "author_url": "",
      "post_date": "09/02/2020 11:12:00",
      "content": "<p>Well, I think it's worth a try. We do it all the time and it definitely diversifies the audio data. You might wanna think about a more elaborate way of doing this (e.g., weighted sum) but mix-up in general should work just fine. </p>",
      "votes": null,
      "replies": [
        {
          "id": 995393,
          "author_name": "quandapro",
          "author_url": "",
          "post_date": "09/02/2020 12:04:16",
          "content": "<p>I definately will give it a shot!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "993961": "I had great result with \"Mix-up\" for supervised computer vision problems. I wonder if \"Mix-up\" can do the same thing for audio.\n```\nmixed_up = a*wav_1 + b*wav_2\nnew_labels = a*label_1 + b*label_2\n```",
    "995333": "Well, I think it's worth a try. We do it all the time and it definitely diversifies the audio data. You might wanna think about a more elaborate way of doing this (e.g., weighted sum) but mix-up in general should work just fine.",
    "995393": "I definately will give it a shot!"
  },
  "source": "meta"
}