{
  "id": 663610,
  "title": "Advice Needed: Should an AI Beginner Join This Competition or Focus on Learning First?",
  "url": "/competitions/deep-past-initiative-machine-translation/discussion/663610",
  "author_name": "",
  "post_date": "2025-12-19T05:50:03.415355300Z",
  "votes": 3,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I am a beginner in the field of artificial intelligence. Currently, I only know how to use PyTorch and some basic NLP processing tools. I would like to ask the seniors whether this competition is suitable for a beginner like me, or if it’s better not to waste time and just focus on learning more courses first. I know this question is a bit silly, but I still want to ask for some advice.</p>",
  "messages": [
    {
      "id": "3379072",
      "postDate": "12/19/2025 05:50:03",
      "content": "<p>I am a beginner in the field of artificial intelligence. Currently, I only know how to use PyTorch and some basic NLP processing tools. I would like to ask the seniors whether this competition is suitable for a beginner like me, or if it’s better not to waste time and just focus on learning more courses first. I know this question is a bit silly, but I still want to ask for some advice.</p>",
      "rawMarkdown": "I am a beginner in the field of artificial intelligence. Currently, I only know how to use PyTorch and some basic NLP processing tools. I would like to ask the seniors whether this competition is suitable for a beginner like me, or if it’s better not to waste time and just focus on learning more courses first. I know this question is a bit silly, but I still want to ask for some advice.",
      "votes": null
    },
    {
      "id": "3379137",
      "postDate": "12/19/2025 08:50:46",
      "content": "<p>I would do the playground series first and then return to this competition, or switch to tutorials if you find you aren't ready</p>",
      "rawMarkdown": "I would do the playground series first and then return to this competition, or switch to tutorials if you find you aren't ready",
      "votes": null
    },
    {
      "id": "3379213",
      "postDate": "12/19/2025 11:26:50",
      "content": "<p>Hi, this is not a silly question at all, and it’s good that you’re asking it early.</p>\n<p>Short answer: yes, a beginner can join, but only if you treat the competition as a learning exercise, not as a performance benchmark.</p>\n<p>This competition is challenging because it combines NLP, low-resource languages, and domain-specific quirks. You should not expect to be competitive on the leaderboard at first, and that’s completely fine. The real value for a beginner is in:</p>\n<ul>\n<li>understanding how a real competition dataset is structured</li>\n<li>learning how baselines are built</li>\n<li>experimenting with preprocessing, tokenization, and evaluation</li>\n<li>reading other people’s notebooks and discussions</li>\n</ul>\n<p>If you wait until you feel “ready,” you may never start. At the same time, don’t abandon structured learning either. The best approach is often both:</p>\n<ul>\n<li>keep learning fundamentals (PyTorch, NLP basics, transformers, etc.)</li>\n<li>participate in the competition alongside that learning, without pressure</li>\n</ul>\n<p>Think of the competition as a sandbox, not an exam. Even a simple baseline submission is a success if you learn something from it.</p>\n<p>If you find it overwhelming, you can step back at any time. Nothing is wasted if you come out understanding a bit more than you did before.</p>",
      "rawMarkdown": "Hi, this is not a silly question at all, and it’s good that you’re asking it early.\n\nShort answer: yes, a beginner can join, but only if you treat the competition as a learning exercise, not as a performance benchmark.\n\nThis competition is challenging because it combines NLP, low-resource languages, and domain-specific quirks. You should not expect to be competitive on the leaderboard at first, and that’s completely fine. The real value for a beginner is in:\n- understanding how a real competition dataset is structured\n- learning how baselines are built\n- experimenting with preprocessing, tokenization, and evaluation\n- reading other people’s notebooks and discussions\n\nIf you wait until you feel “ready,” you may never start. At the same time, don’t abandon structured learning either. The best approach is often both:\n- keep learning fundamentals (PyTorch, NLP basics, transformers, etc.)\n- participate in the competition alongside that learning, without pressure\n\nThink of the competition as a sandbox, not an exam. Even a simple baseline submission is a success if you learn something from it.\n\nIf you find it overwhelming, you can step back at any time. Nothing is wasted if you come out understanding a bit more than you did before.",
      "votes": null
    },
    {
      "id": "3379249",
      "postDate": "12/19/2025 12:55:45",
      "content": "<p>From my experience, attending is a good idea, the score is not important. Just take it as an interesting practice task and learn from public notebooks and the methods of others. I remember when I participated in the first competition, I didn't even know what CNN was.</p>",
      "rawMarkdown": "From my experience, attending is a good idea, the score is not important. Just take it as an interesting practice task and learn from public notebooks and the methods of others. I remember when I participated in the first competition, I didn't even know what CNN was.",
      "votes": null
    },
    {
      "id": "3379829",
      "postDate": "12/20/2025 22:01:07",
      "content": "<p>any nmt tutorial you found helpfull</p>",
      "rawMarkdown": "any nmt tutorial you found helpfull",
      "votes": null
    },
    {
      "id": "3380334",
      "postDate": "12/22/2025 08:48:31",
      "content": "<p>Not NMT directly, but there are many awesome tutorials on NLP with Transformers that are great for grasping the general concepts and how to work with text data.</p>",
      "rawMarkdown": "Not NMT directly, but there are many awesome tutorials on NLP with Transformers that are great for grasping the general concepts and how to work with text data.",
      "votes": null
    },
    {
      "id": "3380437",
      "postDate": "12/22/2025 12:21:13",
      "content": "<p>you are right but looking for directly on nmt maybe we could find something usefull to boost our score</p>",
      "rawMarkdown": "you are right but looking for directly on nmt maybe we could find something usefull to boost our score",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3379137,
      "author_name": "milanmuntenjon",
      "author_url": "",
      "post_date": "12/19/2025 08:50:46",
      "content": "<p>I would do the playground series first and then return to this competition, or switch to tutorials if you find you aren't ready</p>",
      "votes": null,
      "replies": [
        {
          "id": 3379829,
          "author_name": "yusufsertkayaysk",
          "author_url": "",
          "post_date": "12/20/2025 22:01:07",
          "content": "<p>any nmt tutorial you found helpfull</p>",
          "votes": null,
          "replies": [
            {
              "id": 3380334,
              "author_name": "milanmuntenjon",
              "author_url": "",
              "post_date": "12/22/2025 08:48:31",
              "content": "<p>Not NMT directly, but there are many awesome tutorials on NLP with Transformers that are great for grasping the general concepts and how to work with text data.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3380437,
                  "author_name": "yusufsertkayaysk",
                  "author_url": "",
                  "post_date": "12/22/2025 12:21:13",
                  "content": "<p>you are right but looking for directly on nmt maybe we could find something usefull to boost our score</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    },
    {
      "id": 3379213,
      "author_name": "aridoge13",
      "author_url": "",
      "post_date": "12/19/2025 11:26:50",
      "content": "<p>Hi, this is not a silly question at all, and it’s good that you’re asking it early.</p>\n<p>Short answer: yes, a beginner can join, but only if you treat the competition as a learning exercise, not as a performance benchmark.</p>\n<p>This competition is challenging because it combines NLP, low-resource languages, and domain-specific quirks. You should not expect to be competitive on the leaderboard at first, and that’s completely fine. The real value for a beginner is in:</p>\n<ul>\n<li>understanding how a real competition dataset is structured</li>\n<li>learning how baselines are built</li>\n<li>experimenting with preprocessing, tokenization, and evaluation</li>\n<li>reading other people’s notebooks and discussions</li>\n</ul>\n<p>If you wait until you feel “ready,” you may never start. At the same time, don’t abandon structured learning either. The best approach is often both:</p>\n<ul>\n<li>keep learning fundamentals (PyTorch, NLP basics, transformers, etc.)</li>\n<li>participate in the competition alongside that learning, without pressure</li>\n</ul>\n<p>Think of the competition as a sandbox, not an exam. Even a simple baseline submission is a success if you learn something from it.</p>\n<p>If you find it overwhelming, you can step back at any time. Nothing is wasted if you come out understanding a bit more than you did before.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3379249,
      "author_name": "horikitasaku",
      "author_url": "",
      "post_date": "12/19/2025 12:55:45",
      "content": "<p>From my experience, attending is a good idea, the score is not important. Just take it as an interesting practice task and learn from public notebooks and the methods of others. I remember when I participated in the first competition, I didn't even know what CNN was.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3379072": "I am a beginner in the field of artificial intelligence. Currently, I only know how to use PyTorch and some basic NLP processing tools. I would like to ask the seniors whether this competition is suitable for a beginner like me, or if it’s better not to waste time and just focus on learning more courses first. I know this question is a bit silly, but I still want to ask for some advice.",
    "3379137": "I would do the playground series first and then return to this competition, or switch to tutorials if you find you aren't ready",
    "3379213": "Hi, this is not a silly question at all, and it’s good that you’re asking it early.\n\nShort answer: yes, a beginner can join, but only if you treat the competition as a learning exercise, not as a performance benchmark.\n\nThis competition is challenging because it combines NLP, low-resource languages, and domain-specific quirks. You should not expect to be competitive on the leaderboard at first, and that’s completely fine. The real value for a beginner is in:\n- understanding how a real competition dataset is structured\n- learning how baselines are built\n- experimenting with preprocessing, tokenization, and evaluation\n- reading other people’s notebooks and discussions\n\nIf you wait until you feel “ready,” you may never start. At the same time, don’t abandon structured learning either. The best approach is often both:\n- keep learning fundamentals (PyTorch, NLP basics, transformers, etc.)\n- participate in the competition alongside that learning, without pressure\n\nThink of the competition as a sandbox, not an exam. Even a simple baseline submission is a success if you learn something from it.\n\nIf you find it overwhelming, you can step back at any time. Nothing is wasted if you come out understanding a bit more than you did before.",
    "3379249": "From my experience, attending is a good idea, the score is not important. Just take it as an interesting practice task and learn from public notebooks and the methods of others. I remember when I participated in the first competition, I didn't even know what CNN was.",
    "3379829": "any nmt tutorial you found helpfull",
    "3380334": "Not NMT directly, but there are many awesome tutorials on NLP with Transformers that are great for grasping the general concepts and how to work with text data.",
    "3380437": "you are right but looking for directly on nmt maybe we could find something usefull to boost our score"
  },
  "source": "meta"
}