As i've remarked, the majority of competitors uses research papers to build good models.\nI've found that most of the papers come from:\n- [0]: https://arxiv.org/\n- [1]: https://iclr.cc/\n- [2]: https://papers.nips.cc/\n- [3]: https://icml.cc/\n- [4]: http://openaccess.thecvf.com/CVPR2018.py\n- [5]: https://www.cs.toronto.edu/~hinton/papers.html\n- [6]: http://yann.lecun.com/exdb/publis/index.html\n- [.]: https://www.ieee.org/publications/index.html\nIf you know any others, please add theme in a comment.
", "messages": [ { "id": "456443", "postDate": "01/15/2019 20:30:52", "content": "As i've remarked, the majority of competitors uses research papers to build good models.\nI've found that most of the papers come from:\n- [0]: https://arxiv.org/\n- [1]: https://iclr.cc/\n- [2]: https://papers.nips.cc/\n- [3]: https://icml.cc/\n- [4]: http://openaccess.thecvf.com/CVPR2018.py\n- [5]: https://www.cs.toronto.edu/~hinton/papers.html\n- [6]: http://yann.lecun.com/exdb/publis/index.html\n- [.]: https://www.ieee.org/publications/index.html\nIf you know any others, please add theme in a comment.
", "rawMarkdown": "As i've remarked, the majority of competitors uses research papers to build good models.\nI've found that most of the papers come from:\n- [0]: https://arxiv.org/\n- [1]: https://iclr.cc/\n- [2]: https://papers.nips.cc/\n- [3]: https://icml.cc/\n- [4]: http://openaccess.thecvf.com/CVPR2018.py\n- [5]: https://www.cs.toronto.edu/~hinton/papers.html\n- [6]: http://yann.lecun.com/exdb/publis/index.html\n- [.]: https://www.ieee.org/publications/index.html\nIf you know any others, please add theme in a comment.", "votes": null }, { "id": "456477", "postDate": "01/15/2019 22:24:21", "content": "For NLP, ACL anthology is the most famous host.\nhttps://aclanthology.info/
", "rawMarkdown": "For NLP, ACL anthology is the most famous host.\nhttps://aclanthology.info/", "votes": null }, { "id": "456556", "postDate": "01/16/2019 03:35:34", "content": "Um, there's way too many papers written every month. Virtually every ML paper is hosted on Arxiv, but there's really no point in looking for ideas there. There's just too many, most of them are out of date and/or irrelevant. Worst of all, all of them are hard to reproduce! I think you will have much more success by checking out ML forums, or simply looking at different kernels from this and similar competitions.
\n\nReading the currently hyped papers is still fun and I would recommend it just to stay in the loop. They might give you ideas for future competitions, too! The best way to find them is probably via Arxiv-Sanity
\n\nGood luck!
", "rawMarkdown": "Um, there's way too many papers written every month. Virtually every ML paper is hosted on Arxiv, but there's really no point in looking for ideas there. There's just too many, most of them are out of date and/or irrelevant. Worst of all, all of them are hard to reproduce! I think you will have much more success by checking out ML forums, or simply looking at different kernels from this and similar competitions.\n\nReading the currently hyped papers is still fun and I would recommend it just to stay in the loop. They might give you ideas for future competitions, too! The best way to find them is probably via [Arxiv-Sanity](http://www.arxiv-sanity.com/top)\n\nGood luck!", "votes": null } ], "comments": [ { "id": 456477, "author_name": "yufuin", "author_url": "", "post_date": "01/15/2019 22:24:21", "content": "For NLP, ACL anthology is the most famous host.\nhttps://aclanthology.info/
", "votes": null, "replies": [] }, { "id": 456556, "author_name": "mschumacher", "author_url": "", "post_date": "01/16/2019 03:35:34", "content": "Um, there's way too many papers written every month. Virtually every ML paper is hosted on Arxiv, but there's really no point in looking for ideas there. There's just too many, most of them are out of date and/or irrelevant. Worst of all, all of them are hard to reproduce! I think you will have much more success by checking out ML forums, or simply looking at different kernels from this and similar competitions.
\n\nReading the currently hyped papers is still fun and I would recommend it just to stay in the loop. They might give you ideas for future competitions, too! The best way to find them is probably via Arxiv-Sanity
\n\nGood luck!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "456443": "As i've remarked, the majority of competitors uses research papers to build good models.\nI've found that most of the papers come from:\n- [0]: https://arxiv.org/\n- [1]: https://iclr.cc/\n- [2]: https://papers.nips.cc/\n- [3]: https://icml.cc/\n- [4]: http://openaccess.thecvf.com/CVPR2018.py\n- [5]: https://www.cs.toronto.edu/~hinton/papers.html\n- [6]: http://yann.lecun.com/exdb/publis/index.html\n- [.]: https://www.ieee.org/publications/index.html\nIf you know any others, please add theme in a comment.", "456477": "For NLP, ACL anthology is the most famous host.\nhttps://aclanthology.info/", "456556": "Um, there's way too many papers written every month. Virtually every ML paper is hosted on Arxiv, but there's really no point in looking for ideas there. There's just too many, most of them are out of date and/or irrelevant. Worst of all, all of them are hard to reproduce! I think you will have much more success by checking out ML forums, or simply looking at different kernels from this and similar competitions.\n\nReading the currently hyped papers is still fun and I would recommend it just to stay in the loop. They might give you ideas for future competitions, too! The best way to find them is probably via [Arxiv-Sanity](http://www.arxiv-sanity.com/top)\n\nGood luck!" }, "source": "meta" }