{
  "id": 322344,
  "title": "Are you using Few-shot learning?",
  "url": "/competitions/birdclef-2022/discussion/322344",
  "author_name": "",
  "post_date": "2022-05-01T15:45:35.093777700Z",
  "votes": 4,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Are any participants trying to use Few-shot learning to train a main classifier model?<br>\nIf you don't mind me asking, what is the strategy for Few-shot learning that you are trying to use?<br>\nBy the way, I am considering submitting with a prototypical network first.</p>",
  "messages": [
    {
      "id": "1773927",
      "postDate": "05/01/2022 15:45:35",
      "content": "<p>Are any participants trying to use Few-shot learning to train a main classifier model?<br>\nIf you don't mind me asking, what is the strategy for Few-shot learning that you are trying to use?<br>\nBy the way, I am considering submitting with a prototypical network first.</p>",
      "rawMarkdown": "Are any participants trying to use Few-shot learning to train a main classifier model?\nIf you don't mind me asking, what is the strategy for Few-shot learning that you are trying to use?\nBy the way, I am considering submitting with a prototypical network first.",
      "votes": null
    },
    {
      "id": "1774646",
      "postDate": "05/02/2022 11:02:35",
      "content": "<p>This comp (and future) involves few-shot learning more and more. It would be good to find out approaches for the same. As of now yours is the only published notebook on this topic. I wish to do more research in this area but may not get the time. </p>",
      "rawMarkdown": "This comp (and future) involves few-shot learning more and more. It would be good to find out approaches for the same. As of now yours is the only published notebook on this topic. I wish to do more research in this area but may not get the time.",
      "votes": null
    },
    {
      "id": "1774652",
      "postDate": "05/02/2022 11:14:05",
      "content": "<p>Yeah, I agree.<br>\nNow I looking back, I think I underestimated about the cost adapting a completely new approach to the real problem. I read a lot of paper, but it became so close to the deadline and I doubt I can adopt all new ideas.</p>\n<p>Thanks for the comment. I was thinking no one is interested in this topic.</p>",
      "rawMarkdown": "Yeah, I agree.\nNow I looking back, I think I underestimated about the cost adapting a completely new approach to the real problem. I read a lot of paper, but it became so close to the deadline and I doubt I can adopt all new ideas.\n\nThanks for the comment. I was thinking no one is interested in this topic.",
      "votes": null
    },
    {
      "id": "1774688",
      "postDate": "05/02/2022 11:35:56",
      "content": "<p>I am sure the knowledge you gained will come of use in subsequent comps. Definitely few shot learning is here to stay I feel. There may be a few folks using various fewshot strategies in this comp itself. I suppose they will share it post the competition ends.</p>",
      "rawMarkdown": "I am sure the knowledge you gained will come of use in subsequent comps. Definitely few shot learning is here to stay I feel. There may be a few folks using various fewshot strategies in this comp itself. I suppose they will share it post the competition ends.",
      "votes": null
    },
    {
      "id": "1774740",
      "postDate": "05/02/2022 12:09:30",
      "content": "<p>Thanks. And for me, it isn’t too late to adopt some of my idea too.</p>",
      "rawMarkdown": "Thanks. And for me, it isn’t too late to adopt some of my idea too.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1774646,
      "author_name": "allohvk",
      "author_url": "",
      "post_date": "05/02/2022 11:02:35",
      "content": "<p>This comp (and future) involves few-shot learning more and more. It would be good to find out approaches for the same. As of now yours is the only published notebook on this topic. I wish to do more research in this area but may not get the time. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1774652,
          "author_name": "tatamikenn",
          "author_url": "",
          "post_date": "05/02/2022 11:14:05",
          "content": "<p>Yeah, I agree.<br>\nNow I looking back, I think I underestimated about the cost adapting a completely new approach to the real problem. I read a lot of paper, but it became so close to the deadline and I doubt I can adopt all new ideas.</p>\n<p>Thanks for the comment. I was thinking no one is interested in this topic.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1774688,
          "author_name": "allohvk",
          "author_url": "",
          "post_date": "05/02/2022 11:35:56",
          "content": "<p>I am sure the knowledge you gained will come of use in subsequent comps. Definitely few shot learning is here to stay I feel. There may be a few folks using various fewshot strategies in this comp itself. I suppose they will share it post the competition ends.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1774740,
          "author_name": "tatamikenn",
          "author_url": "",
          "post_date": "05/02/2022 12:09:30",
          "content": "<p>Thanks. And for me, it isn’t too late to adopt some of my idea too.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1773927": "Are any participants trying to use Few-shot learning to train a main classifier model?\nIf you don't mind me asking, what is the strategy for Few-shot learning that you are trying to use?\nBy the way, I am considering submitting with a prototypical network first.",
    "1774646": "This comp (and future) involves few-shot learning more and more. It would be good to find out approaches for the same. As of now yours is the only published notebook on this topic. I wish to do more research in this area but may not get the time.",
    "1774652": "Yeah, I agree.\nNow I looking back, I think I underestimated about the cost adapting a completely new approach to the real problem. I read a lot of paper, but it became so close to the deadline and I doubt I can adopt all new ideas.\n\nThanks for the comment. I was thinking no one is interested in this topic.",
    "1774688": "I am sure the knowledge you gained will come of use in subsequent comps. Definitely few shot learning is here to stay I feel. There may be a few folks using various fewshot strategies in this comp itself. I suppose they will share it post the competition ends.",
    "1774740": "Thanks. And for me, it isn’t too late to adopt some of my idea too."
  },
  "source": "meta"
}