{
  "id": 223381,
  "title": "Winning solutions from similar competition: Molecular Translation into SMILES",
  "url": "/competitions/bms-molecular-translation/discussion/223381",
  "author_name": "",
  "post_date": "2021-03-03T15:18:40.715149600Z",
  "votes": 100,
  "comment_count": 29,
  "views": 0,
  "content": "<p>Last year, similar competition has been held on <a href=\"https://dacon.io/\" target=\"_blank\">DACON</a>.</p>\n<p><a href=\"https://www.kaggle.com/jinssaa\" target=\"_blank\">@jinssaa</a> has kindly shared the links to the solution PPT on this <a href=\"https://www.kaggle.com/c/bms-molecular-translation/discussion/223268\" target=\"_blank\">discussion</a></p>\n<p>In this post, I'm going to summarize the winning solutions on that competition in English which could provide some insights for this competition. </p>\n<h2>Competition Overview</h2>\n<ul>\n<li>Purpose: Molecular Translation into SMILES</li>\n<li>Number of provided images for training: about 1M</li>\n<li>Evaluation: Tanimoto Similarity</li>\n</ul>\n<h4>UPDATE MARCH 26, 2021</h4>\n<ul>\n<li>Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions</li>\n<li>It seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers</li>\n<li>I will share it again when the issue resolves </li>\n</ul>\n<h4>UPDATE MARCH 29, 2021</h4>\n<ul>\n<li>I was informed that the decision has been finalized as to not share the solutions. </li>\n<li>I apologize to all the new comers into this post expecting to see the solutions. </li>\n</ul>",
  "messages": [
    {
      "id": "1225374",
      "postDate": "03/03/2021 15:18:40",
      "content": "<p>Last year, similar competition has been held on <a href=\"https://dacon.io/\" target=\"_blank\">DACON</a>.</p>\n<p><a href=\"https://www.kaggle.com/jinssaa\" target=\"_blank\">@jinssaa</a> has kindly shared the links to the solution PPT on this <a href=\"https://www.kaggle.com/c/bms-molecular-translation/discussion/223268\" target=\"_blank\">discussion</a></p>\n<p>In this post, I'm going to summarize the winning solutions on that competition in English which could provide some insights for this competition. </p>\n<h2>Competition Overview</h2>\n<ul>\n<li>Purpose: Molecular Translation into SMILES</li>\n<li>Number of provided images for training: about 1M</li>\n<li>Evaluation: Tanimoto Similarity</li>\n</ul>\n<h4>UPDATE MARCH 26, 2021</h4>\n<ul>\n<li>Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions</li>\n<li>It seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers</li>\n<li>I will share it again when the issue resolves </li>\n</ul>\n<h4>UPDATE MARCH 29, 2021</h4>\n<ul>\n<li>I was informed that the decision has been finalized as to not share the solutions. </li>\n<li>I apologize to all the new comers into this post expecting to see the solutions. </li>\n</ul>",
      "rawMarkdown": "Last year, similar competition has been held on [DACON](https://dacon.io/).\n\n@jinssaa has kindly shared the links to the solution PPT on this [discussion](https://www.kaggle.com/c/bms-molecular-translation/discussion/223268)\n\nIn this post, I'm going to summarize the winning solutions on that competition in English which could provide some insights for this competition. \n\n## Competition Overview\n- Purpose: Molecular Translation into SMILES\n- Number of provided images for training: about 1M\n- Evaluation: Tanimoto Similarity\n\n#### UPDATE MARCH 26, 2021 \n- Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\n- It seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers\n- I will share it again when the issue resolves \n\n#### UPDATE MARCH 29, 2021 \n- I was informed that the decision has been finalized as to not share the solutions. \n- I apologize to all the new comers into this post expecting to see the solutions.",
      "votes": null
    },
    {
      "id": "1225482",
      "postDate": "03/03/2021 16:24:27",
      "content": "<p>I'm curious about the quality of data provided for that competition. For this one the resolution doesn't seem to be too high, in that case, would PubChem data be useful?</p>",
      "rawMarkdown": "I'm curious about the quality of data provided for that competition. For this one the resolution doesn't seem to be too high, in that case, would PubChem data be useful?",
      "votes": null
    },
    {
      "id": "1225495",
      "postDate": "03/03/2021 16:35:55",
      "content": "<p>They provided images in fixed size of 299 x 299 x 3. You can see samples in this <a href=\"https://dacon.io/competitions/official/235640/talkboard/401436?page=3&amp;dtype=recent&amp;ptype=pub\" target=\"_blank\">discussion</a>. The resolution wasn't also that high. </p>\n<p>To the best of my knowledge, I'm not quite sure if PubChem's data can be useful for solving low-resolution problem. I haven't looked at it yet. </p>\n<p>Based on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization. </p>",
      "rawMarkdown": "They provided images in fixed size of 299 x 299 x 3. You can see samples in this [discussion](https://dacon.io/competitions/official/235640/talkboard/401436?page=3&dtype=recent&ptype=pub). The resolution wasn't also that high. \n\nTo the best of my knowledge, I'm not quite sure if PubChem's data can be useful for solving low-resolution problem. I haven't looked at it yet. \n\nBased on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization.",
      "votes": null
    },
    {
      "id": "1225507",
      "postDate": "03/03/2021 16:44:34",
      "content": "<p>Yeah they do look low-res to me, albeit a bit different from the data provided in the current competition.</p>\n<blockquote>\n  <p>Based on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization.</p>\n</blockquote>\n<p>Yes but from what I've observed, PubChem data looks a lot different. I'm excited to try out stuff using that data anyway ;)</p>",
      "rawMarkdown": "Yeah they do look low-res to me, albeit a bit different from the data provided in the current competition.\n\n> Based on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization.\n\nYes but from what I've observed, PubChem data looks a lot different. I'm excited to try out stuff using that data anyway ;)",
      "votes": null
    },
    {
      "id": "1225579",
      "postDate": "03/03/2021 18:06:43",
      "content": "<p>Thank you for posting this information!</p>",
      "rawMarkdown": "Thank you for posting this information!",
      "votes": null
    },
    {
      "id": "1225838",
      "postDate": "03/04/2021 01:57:04",
      "content": "<p>Thank you for this interesting competition!</p>",
      "rawMarkdown": "Thank you for this interesting competition!",
      "votes": null
    },
    {
      "id": "1226761",
      "postDate": "03/04/2021 20:26:25",
      "content": "<p>Thank you for summarizing all the three solutions much helpful</p>",
      "rawMarkdown": "Thank you for summarizing all the three solutions much helpful",
      "votes": null
    },
    {
      "id": "1228081",
      "postDate": "03/06/2021 05:17:27",
      "content": "<p>Thanks for sharing this, very useful!</p>",
      "rawMarkdown": "Thanks for sharing this, very useful!",
      "votes": null
    },
    {
      "id": "1228091",
      "postDate": "03/06/2021 05:31:59",
      "content": "<p>My pleasure!</p>",
      "rawMarkdown": "My pleasure!",
      "votes": null
    },
    {
      "id": "1228093",
      "postDate": "03/06/2021 05:33:31",
      "content": "<p>I'm glad that it helped.</p>",
      "rawMarkdown": "I'm glad that it helped.",
      "votes": null
    },
    {
      "id": "1230403",
      "postDate": "03/08/2021 05:19:23",
      "content": "<p><a href=\"https://www.kaggle.com/jiny333\" target=\"_blank\">@jiny333</a> Thanks alot for sharing and translation :) , This is good starting point for this competition :) .<br>\nIf Possible can you please also add cross validation strategies used by different teams , That would be great :) .</p>",
      "rawMarkdown": "jiny333 Thanks alot for sharing and translation :) , This is good starting point for this competition :) .\nIf Possible can you please also add cross validation strategies used by different teams , That would be great :) .",
      "votes": null
    },
    {
      "id": "1231055",
      "postDate": "03/08/2021 16:30:39",
      "content": "<p><a href=\"https://www.kaggle.com/jiny333\" target=\"_blank\">@jiny333</a> Thanks for sharing the solutions . Looks like \" Transformer is all you need \"</p>",
      "rawMarkdown": "jiny333 Thanks for sharing the solutions . Looks like \" Transformer is all you need \"",
      "votes": null
    },
    {
      "id": "1233685",
      "postDate": "03/10/2021 16:09:48",
      "content": "<p>link forbidden… Can we download those solution ppts from somewhere else?</p>",
      "rawMarkdown": "link forbidden... Can we download those solution ppts from somewhere else?",
      "votes": null
    },
    {
      "id": "1235886",
      "postDate": "03/12/2021 14:56:28",
      "content": "<p>Thanks for sharing. It's time to learn RCNN lol</p>",
      "rawMarkdown": "Thanks for sharing. It's time to learn RCNN lol",
      "votes": null
    },
    {
      "id": "1237347",
      "postDate": "03/14/2021 04:23:34",
      "content": "<p>I'll work on that as soon as time allows. </p>",
      "rawMarkdown": "I'll work on that as soon as time allows.",
      "votes": null
    },
    {
      "id": "1237348",
      "postDate": "03/14/2021 04:23:51",
      "content": "<p>Yep~ Good luck!</p>",
      "rawMarkdown": "Yep~ Good luck!",
      "votes": null
    },
    {
      "id": "1237355",
      "postDate": "03/14/2021 04:25:39",
      "content": "<p>I think <a href=\"https://www.kaggle.com/jinssaa\" target=\"_blank\">@jinssaa</a> has posted the link to slideshare on the comment on this <a href=\"https://www.kaggle.com/c/bms-molecular-translation/discussion/223268\" target=\"_blank\">post</a></p>",
      "rawMarkdown": "I think @jinssaa has posted the link to slideshare on the comment on this [post](https://www.kaggle.com/c/bms-molecular-translation/discussion/223268)",
      "votes": null
    },
    {
      "id": "1237424",
      "postDate": "03/14/2021 06:09:49",
      "content": "<p><a href=\"https://www.kaggle.com/jiny333\" target=\"_blank\">@jiny333</a>  yes it would be very kind of you if you can translate CV strategy as well. :)</p>",
      "rawMarkdown": "jiny333  yes it would be very kind of you if you can translate CV strategy as well. :)",
      "votes": null
    },
    {
      "id": "1239427",
      "postDate": "03/15/2021 17:54:58",
      "content": "<p>Thanks a lot for sharing , did u try to implement one of those solutions ? </p>",
      "rawMarkdown": "Thanks a lot for sharing , did u try to implement one of those solutions ?",
      "votes": null
    },
    {
      "id": "1250517",
      "postDate": "03/24/2021 05:07:37",
      "content": "<p>Thanks a lot. Really informative.</p>",
      "rawMarkdown": "Thanks a lot. Really informative.",
      "votes": null
    },
    {
      "id": "1251195",
      "postDate": "03/24/2021 15:13:18",
      "content": "<p>I haven't yet, I plan on doing so as soon as I get some free time. </p>",
      "rawMarkdown": "I haven't yet, I plan on doing so as soon as I get some free time.",
      "votes": null
    },
    {
      "id": "1251197",
      "postDate": "03/24/2021 15:14:05",
      "content": "<p>My pleasure.</p>",
      "rawMarkdown": "My pleasure.",
      "votes": null
    },
    {
      "id": "1251937",
      "postDate": "03/25/2021 09:12:09",
      "content": "<p>Can we get the data of PubChem?</p>",
      "rawMarkdown": "Can we get the data of PubChem?",
      "votes": null
    },
    {
      "id": "1253174",
      "postDate": "03/26/2021 12:38:52",
      "content": "<p>wow, that's really weird they have done it now. If it was an open source project and they've already shared their code, why deleting it now? I mean open sourcing is exactly for these times in order not to re-invent the wheel and it gives opportunity to improve upon what exists. Sounds really funny to me what they did!</p>",
      "rawMarkdown": "wow, that's really weird they have done it now. If it was an open source project and they've already shared their code, why deleting it now? I mean open sourcing is exactly for these times in order not to re-invent the wheel and it gives opportunity to improve upon what exists. Sounds really funny to me what they did!",
      "votes": null
    },
    {
      "id": "1253215",
      "postDate": "03/26/2021 13:24:52",
      "content": "<p>Thats a huge surprise. Even the original Kaggle post got removed =)</p>\n<p>Edit: it seems like a copyright issue </p>\n<pre><code>Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\nIt seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers\nI will share it again when the issue resolves\n</code></pre>",
      "rawMarkdown": "Thats a huge surprise. Even the original Kaggle post got removed =)\n\nEdit: it seems like a copyright issue \n\n```\nUnfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\nIt seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers\nI will share it again when the issue resolves\n```",
      "votes": null
    },
    {
      "id": "1259498",
      "postDate": "04/01/2021 13:09:21",
      "content": "<p>Wow, this creates a very large unfairness between those who had access to these solutions and those who haven't.</p>",
      "rawMarkdown": "Wow, this creates a very large unfairness between those who had access to these solutions and those who haven't.",
      "votes": null
    },
    {
      "id": "1259501",
      "postDate": "04/01/2021 13:10:50",
      "content": "<blockquote>\n  <p>Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions</p>\n</blockquote>\n<p>Copyright is on the actual content, not on a summary you could have made.  Before complying to a legal order like this better check if there is any ground for it.  </p>",
      "rawMarkdown": "> Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\n\nCopyright is on the actual content, not on a summary you could have made.  Before complying to a legal order like this better check if there is any ground for it.",
      "votes": null
    },
    {
      "id": "1259747",
      "postDate": "04/01/2021 16:24:11",
      "content": "<p>Even a kaggle kernel was deleted. I don't get what that has to do with copyright issue. </p>",
      "rawMarkdown": "Even a kaggle kernel was deleted. I don't get what that has to do with copyright issue.",
      "votes": null
    },
    {
      "id": "1260189",
      "postDate": "04/01/2021 23:18:11",
      "content": "<p>Now that's a hint there. </p>",
      "rawMarkdown": "Now that's a hint there.",
      "votes": null
    },
    {
      "id": "1260509",
      "postDate": "04/02/2021 07:26:15",
      "content": "<p><a href=\"https://www.kaggle.com/serigne\" target=\"_blank\">@serigne</a> If the kernel included code that was copied from a copyrighted source the there is a legal liability.  I don't know if it was the case for the kernel you mention though.</p>",
      "rawMarkdown": "serigne If the kernel included code that was copied from a copyrighted source the there is a legal liability.  I don't know if it was the case for the kernel you mention though.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1225482,
      "author_name": "arka47",
      "author_url": "",
      "post_date": "03/03/2021 16:24:27",
      "content": "<p>I'm curious about the quality of data provided for that competition. For this one the resolution doesn't seem to be too high, in that case, would PubChem data be useful?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1225495,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/03/2021 16:35:55",
          "content": "<p>They provided images in fixed size of 299 x 299 x 3. You can see samples in this <a href=\"https://dacon.io/competitions/official/235640/talkboard/401436?page=3&amp;dtype=recent&amp;ptype=pub\" target=\"_blank\">discussion</a>. The resolution wasn't also that high. </p>\n<p>To the best of my knowledge, I'm not quite sure if PubChem's data can be useful for solving low-resolution problem. I haven't looked at it yet. </p>\n<p>Based on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1225507,
          "author_name": "arka47",
          "author_url": "",
          "post_date": "03/03/2021 16:44:34",
          "content": "<p>Yeah they do look low-res to me, albeit a bit different from the data provided in the current competition.</p>\n<blockquote>\n  <p>Based on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization.</p>\n</blockquote>\n<p>Yes but from what I've observed, PubChem data looks a lot different. I'm excited to try out stuff using that data anyway ;)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1225579,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "03/03/2021 18:06:43",
      "content": "<p>Thank you for posting this information!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1225838,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/04/2021 01:57:04",
          "content": "<p>Thank you for this interesting competition!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1226761,
      "author_name": "ramjidoolla",
      "author_url": "",
      "post_date": "03/04/2021 20:26:25",
      "content": "<p>Thank you for summarizing all the three solutions much helpful</p>",
      "votes": null,
      "replies": [
        {
          "id": 1228093,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/06/2021 05:33:31",
          "content": "<p>I'm glad that it helped.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1228081,
      "author_name": "saurabhbagchi",
      "author_url": "",
      "post_date": "03/06/2021 05:17:27",
      "content": "<p>Thanks for sharing this, very useful!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1228091,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/06/2021 05:31:59",
          "content": "<p>My pleasure!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1230403,
      "author_name": "sayedathar11",
      "author_url": "",
      "post_date": "03/08/2021 05:19:23",
      "content": "<p><a href=\"https://www.kaggle.com/jiny333\" target=\"_blank\">@jiny333</a> Thanks alot for sharing and translation :) , This is good starting point for this competition :) .<br>\nIf Possible can you please also add cross validation strategies used by different teams , That would be great :) .</p>",
      "votes": null,
      "replies": [
        {
          "id": 1237347,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/14/2021 04:23:34",
          "content": "<p>I'll work on that as soon as time allows. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1237424,
          "author_name": "sayedathar11",
          "author_url": "",
          "post_date": "03/14/2021 06:09:49",
          "content": "<p><a href=\"https://www.kaggle.com/jiny333\" target=\"_blank\">@jiny333</a>  yes it would be very kind of you if you can translate CV strategy as well. :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1231055,
      "author_name": "usharengaraju",
      "author_url": "",
      "post_date": "03/08/2021 16:30:39",
      "content": "<p><a href=\"https://www.kaggle.com/jiny333\" target=\"_blank\">@jiny333</a> Thanks for sharing the solutions . Looks like \" Transformer is all you need \"</p>",
      "votes": null,
      "replies": [
        {
          "id": 1260189,
          "author_name": "ayuraj",
          "author_url": "",
          "post_date": "04/01/2021 23:18:11",
          "content": "<p>Now that's a hint there. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1233685,
      "author_name": "njumario",
      "author_url": "",
      "post_date": "03/10/2021 16:09:48",
      "content": "<p>link forbidden… Can we download those solution ppts from somewhere else?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1237355,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/14/2021 04:25:39",
          "content": "<p>I think <a href=\"https://www.kaggle.com/jinssaa\" target=\"_blank\">@jinssaa</a> has posted the link to slideshare on the comment on this <a href=\"https://www.kaggle.com/c/bms-molecular-translation/discussion/223268\" target=\"_blank\">post</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1235886,
      "author_name": "phykn88",
      "author_url": "",
      "post_date": "03/12/2021 14:56:28",
      "content": "<p>Thanks for sharing. It's time to learn RCNN lol</p>",
      "votes": null,
      "replies": [
        {
          "id": 1237348,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/14/2021 04:23:51",
          "content": "<p>Yep~ Good luck!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1239427,
      "author_name": "salimkhazem",
      "author_url": "",
      "post_date": "03/15/2021 17:54:58",
      "content": "<p>Thanks a lot for sharing , did u try to implement one of those solutions ? </p>",
      "votes": null,
      "replies": [
        {
          "id": 1251195,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/24/2021 15:13:18",
          "content": "<p>I haven't yet, I plan on doing so as soon as I get some free time. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1250517,
      "author_name": "moeinshariatnia",
      "author_url": "",
      "post_date": "03/24/2021 05:07:37",
      "content": "<p>Thanks a lot. Really informative.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1251197,
          "author_name": "jiny333",
          "author_url": "",
          "post_date": "03/24/2021 15:14:05",
          "content": "<p>My pleasure.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1251937,
      "author_name": "dldmw579",
      "author_url": "",
      "post_date": "03/25/2021 09:12:09",
      "content": "<p>Can we get the data of PubChem?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1253174,
      "author_name": "moeinshariatnia",
      "author_url": "",
      "post_date": "03/26/2021 12:38:52",
      "content": "<p>wow, that's really weird they have done it now. If it was an open source project and they've already shared their code, why deleting it now? I mean open sourcing is exactly for these times in order not to re-invent the wheel and it gives opportunity to improve upon what exists. Sounds really funny to me what they did!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1253215,
          "author_name": "drhabib",
          "author_url": "",
          "post_date": "03/26/2021 13:24:52",
          "content": "<p>Thats a huge surprise. Even the original Kaggle post got removed =)</p>\n<p>Edit: it seems like a copyright issue </p>\n<pre><code>Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\nIt seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers\nI will share it again when the issue resolves\n</code></pre>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1259498,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "04/01/2021 13:09:21",
      "content": "<p>Wow, this creates a very large unfairness between those who had access to these solutions and those who haven't.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1259501,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "04/01/2021 13:10:50",
      "content": "<blockquote>\n  <p>Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions</p>\n</blockquote>\n<p>Copyright is on the actual content, not on a summary you could have made.  Before complying to a legal order like this better check if there is any ground for it.  </p>",
      "votes": null,
      "replies": [
        {
          "id": 1259747,
          "author_name": "serigne",
          "author_url": "",
          "post_date": "04/01/2021 16:24:11",
          "content": "<p>Even a kaggle kernel was deleted. I don't get what that has to do with copyright issue. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1260509,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "04/02/2021 07:26:15",
          "content": "<p><a href=\"https://www.kaggle.com/serigne\" target=\"_blank\">@serigne</a> If the kernel included code that was copied from a copyrighted source the there is a legal liability.  I don't know if it was the case for the kernel you mention though.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1225374": "Last year, similar competition has been held on [DACON](https://dacon.io/).\n\n@jinssaa has kindly shared the links to the solution PPT on this [discussion](https://www.kaggle.com/c/bms-molecular-translation/discussion/223268)\n\nIn this post, I'm going to summarize the winning solutions on that competition in English which could provide some insights for this competition. \n\n## Competition Overview\n- Purpose: Molecular Translation into SMILES\n- Number of provided images for training: about 1M\n- Evaluation: Tanimoto Similarity\n\n#### UPDATE MARCH 26, 2021 \n- Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\n- It seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers\n- I will share it again when the issue resolves \n\n#### UPDATE MARCH 29, 2021 \n- I was informed that the decision has been finalized as to not share the solutions. \n- I apologize to all the new comers into this post expecting to see the solutions.",
    "1225482": "I'm curious about the quality of data provided for that competition. For this one the resolution doesn't seem to be too high, in that case, would PubChem data be useful?",
    "1225495": "They provided images in fixed size of 299 x 299 x 3. You can see samples in this [discussion](https://dacon.io/competitions/official/235640/talkboard/401436?page=3&dtype=recent&ptype=pub). The resolution wasn't also that high. \n\nTo the best of my knowledge, I'm not quite sure if PubChem's data can be useful for solving low-resolution problem. I haven't looked at it yet. \n\nBased on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization.",
    "1225507": "Yeah they do look low-res to me, albeit a bit different from the data provided in the current competition.\n\n> Based on the solution slides, 2nd and 3rd place winners used them in order to let the model observe lots of different cases for generalization.\n\nYes but from what I've observed, PubChem data looks a lot different. I'm excited to try out stuff using that data anyway ;)",
    "1225579": "Thank you for posting this information!",
    "1225838": "Thank you for this interesting competition!",
    "1226761": "Thank you for summarizing all the three solutions much helpful",
    "1228081": "Thanks for sharing this, very useful!",
    "1228091": "My pleasure!",
    "1228093": "I'm glad that it helped.",
    "1230403": "jiny333 Thanks alot for sharing and translation :) , This is good starting point for this competition :) .\nIf Possible can you please also add cross validation strategies used by different teams , That would be great :) .",
    "1231055": "jiny333 Thanks for sharing the solutions . Looks like \" Transformer is all you need \"",
    "1233685": "link forbidden... Can we download those solution ppts from somewhere else?",
    "1235886": "Thanks for sharing. It's time to learn RCNN lol",
    "1237347": "I'll work on that as soon as time allows.",
    "1237348": "Yep~ Good luck!",
    "1237355": "I think @jinssaa has posted the link to slideshare on the comment on this [post](https://www.kaggle.com/c/bms-molecular-translation/discussion/223268)",
    "1237424": "jiny333  yes it would be very kind of you if you can translate CV strategy as well. :)",
    "1239427": "Thanks a lot for sharing , did u try to implement one of those solutions ?",
    "1250517": "Thanks a lot. Really informative.",
    "1251195": "I haven't yet, I plan on doing so as soon as I get some free time.",
    "1251197": "My pleasure.",
    "1251937": "Can we get the data of PubChem?",
    "1253174": "wow, that's really weird they have done it now. If it was an open source project and they've already shared their code, why deleting it now? I mean open sourcing is exactly for these times in order not to re-invent the wheel and it gives opportunity to improve upon what exists. Sounds really funny to me what they did!",
    "1253215": "Thats a huge surprise. Even the original Kaggle post got removed =)\n\nEdit: it seems like a copyright issue \n\n```\nUnfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\nIt seems all the contents that I had referred to, which were shared to the public at the time I was writing this post, had been deleted as well by the organizers\nI will share it again when the issue resolves\n```",
    "1259498": "Wow, this creates a very large unfairness between those who had access to these solutions and those who haven't.",
    "1259501": "> Unfortunately, due to copyright issues raised by the organizers of this competition, I was told to delete the solutions\n\nCopyright is on the actual content, not on a summary you could have made.  Before complying to a legal order like this better check if there is any ground for it.",
    "1259747": "Even a kaggle kernel was deleted. I don't get what that has to do with copyright issue.",
    "1260189": "Now that's a hint there.",
    "1260509": "serigne If the kernel included code that was copied from a copyrighted source the there is a legal liability.  I don't know if it was the case for the kernel you mention though."
  },
  "source": "meta"
}