I would like to thank the competition organizer and Kaggle team for hosting this competition.
\nAs you can see from the leaderboard, I think it is a good idea to add ranking points and count towards tiers to increase the activity in this competition.
", "messages": [ { "id": 1707554, "postDate": "2022-02-28T15:13:29.843Z", "content": "I would like to thank the competition organizer and Kaggle team for hosting this competition.
\nAs you can see from the leaderboard, I think it is a good idea to add ranking points and count towards tiers to increase the activity in this competition.
", "rawMarkdown": "I would like to thank the competition organizer and Kaggle team for hosting this competition.\n\nAs you can see from the leaderboard, I think it is a good idea to add ranking points and count towards tiers to increase the activity in this competition.\n\n\n", "votes": 8 }, { "id": 1709946, "postDate": "2022-03-02T16:12:10.903Z", "content": "@faisalalsrheed, thanks for sharing your thoughts. We understand that ranking points would motivate more kagglers to the competition. We also want to make sure that your time invested in this competition is worthwhile. Nonetheless, as a part of the FGVC workshop, our competition is more focused on the workshop itself than the Kaggle system. Let me see if I can pass this on to the workshop organizers, but please be aware that no previous FGVC workshop competitions have been involved in competition tiers.
", "rawMarkdown": "@faisalalsrheed, thanks for sharing your thoughts. We understand that ranking points would motivate more kagglers to the competition. We also want to make sure that your time invested in this competition is worthwhile. Nonetheless, as a part of the FGVC workshop, our competition is more focused on the workshop itself than the Kaggle system. Let me see if I can pass this on to the workshop organizers, but please be aware that no previous FGVC workshop competitions have been involved in competition tiers. ", "votes": 1, "replies": [ { "id": 1709977, "postDate": "2022-03-02T16:49:50.193Z", "content": "Thank you for the clarification.
", "rawMarkdown": "Thank you for the clarification.", "votes": 1 }, { "id": 1735114, "postDate": "2022-03-25T21:38:21.287Z", "content": "Hi John. Thank you for the clarification. I know this is might be out of your control, but I just want to point out that a lot of Kagglers might not be going to be actively pursuing research or attending CVPR workshops not because they are not interested, but more because being an academic is a geographic/financial privilege.
\nWhereas a lot of folks pursuing PhDs have formal education in CS or related fields, and are from famous universities in US/EU or in prestigious national schools (e.g. IITs, PKU, SNU, etc.). Those students had the chance to work on research submitted to top conferences before even starting their PhD, and many started preparing for academia in middle school with tutoring and mentors.
\nOn the other hand, a lot of Kagglers are self-taught and rely on the free compute (free GPU and TPU from Kaggle, Colab, etc) and financial incentives (prizes from competition) to gain skills in ML. Although they might have original ideas that could benefit the research community, they might not have the guidance to work on those. Furthermore, even if they were given the chance to attend a prestigious masters abroad, they might not be able to afford it (most scholarships only cover a part of the costs, not all).
\nI think such workshops organized through Kaggle could be a good way of introducing eager students on Kaggle to academic research from throughout the world. It's easy to think that a lot of Kagglers are ML practitioners that are more interested in engineering than graduate school or academia. The truth is, however, that 43% of those who answered the 2019 survey intend to, or have attained a Master's degree, which is a research-based program in most parts of the world.
\nCompared to CVPR, Kaggle has a much more diverse user base because of the ease of access. In the 2019 survey, Parul showed in her notebook over 150 women from various African countries participating in the survey (63 from Nigeria, 29 from Kenya, 27 from Egypt) compared to 624 women in USA (and 99 in Canada). If you look at the same stats at CVPR19, only 0.2% of the submissions come from any African country, and there's only 1 attendee from Egypt, and 0 from Kenya or Nigeria. Similar observations can be made with the fact that India and Brazil are underrepresented in the CVPR survey vs Kaggle, etc.
\nI know bringing those changes would require herculean efforts. I know CVPR is working hard to improve access to underrepresented communities and those with financial barriers. I just hope that this starts a dialogue on how annual competitions by CVPR could be beneficial to more than just researchers already attending CVPR but also the broader data science community, many of whom are definitely interested in research but simply do not have the capability or incentive to participate.
", "rawMarkdown": "Hi John. Thank you for the clarification. I know this is might be out of your control, but I just want to point out that a lot of Kagglers might not be going to be actively pursuing research or attending CVPR workshops not because they are not interested, but more because being an academic is a geographic/financial privilege. \n\nWhereas a lot of folks pursuing PhDs have formal education in CS or related fields, and are from famous universities in US/EU or in prestigious national schools (e.g. IITs, PKU, SNU, etc.). Those students had the chance to work on research submitted to top conferences before even starting their PhD, and many started preparing for academia in middle school with tutoring and mentors.\n\nOn the other hand, a lot of Kagglers are self-taught and rely on the free compute (free GPU and TPU from Kaggle, Colab, etc) and financial incentives (prizes from competition) to gain skills in ML. Although they might have original ideas that could benefit the research community, they might not have the guidance to work on those. Furthermore, even if they were given the chance to attend a prestigious masters abroad, they might not be able to afford it (most scholarships only cover a part of the costs, not all).\n\nI think such workshops organized through Kaggle could be a good way of introducing eager students on Kaggle to academic research from throughout the world. It's easy to think that a lot of Kagglers are ML practitioners that are more interested in engineering than graduate school or academia. The truth is, however, that 43% of those who answered the [2019 survey](https://www.kaggle.com/competitions/kaggle-survey-2019/data) intend to, or have attained a Master's degree, which is a research-based program in most parts of the world.\n\nCompared to CVPR, Kaggle has a much more diverse user base because of the ease of access. In the 2019 survey, [Parul showed in her notebook](https://www.kaggle.com/code/parulpandey/geek-girls-rising-myth-or-reality) over 150 women from various African countries participating in the survey (63 from Nigeria, 29 from Kenya, 27 from Egypt) compared to 624 women in USA (and 99 in Canada). If you look at the [same stats at CVPR19](https://cvpr2019.thecvf.com/files/CVPR%202019%20-%20Welcome%20Slides%20Final.pdf), only 0.2% of the submissions come from any African country, and there's only 1 attendee from Egypt, and 0 from Kenya or Nigeria. Similar observations can be made with the fact that India and Brazil are underrepresented in the CVPR survey vs Kaggle, etc.\n\nI know bringing those changes would require herculean efforts. I know CVPR is working hard to improve access to underrepresented communities and those with financial barriers. I just hope that this starts a dialogue on how annual competitions by CVPR could be beneficial to more than just *researchers already attending CVPR* but also the broader data science community, many of whom are definitely interested in research but simply do not have the capability or incentive to participate.", "votes": 3 }, { "id": 1735274, "postDate": "2022-03-26T04:21:04.830Z", "content": "On a slightly silly note:
\nMaybe I can help create an unofficial Kaggle x Chai conference where we host most interesting comp winners and solutions. A virtual conference but for Kagglers where we can celebrate the awesome knowledge and Kagglers in the community. 🤔
\nThe podcast is an attempt at the same but its a broadcasted medium and not a discussion format.
", "rawMarkdown": "On a slightly silly note: \n\nMaybe I can help create an unofficial Kaggle x Chai conference where we host most interesting comp winners and solutions. A virtual conference but for Kagglers where we can celebrate the awesome knowledge and Kagglers in the community. 🤔\n\nThe podcast is an attempt at the same but its a broadcasted medium and not a discussion format.", "votes": 1 }, { "id": 1735970, "postDate": "2022-03-26T18:54:08.500Z", "content": "@init27 I don't think this is a silly idea; it's great in fact. I think this sort of exposure could definitely encourage people to participate in otherwise no-prize/points competitions.
", "rawMarkdown": "@init27 I don't think this is a silly idea; it's great in fact. I think this sort of exposure could definitely encourage people to participate in otherwise no-prize/points competitions." } ] } ], "comments": [ { "id": 1709946, "author_name": "John Park", "author_url": "", "post_date": "2022-03-02T16:12:10.903000", "content": "@faisalalsrheed, thanks for sharing your thoughts. We understand that ranking points would motivate more kagglers to the competition. We also want to make sure that your time invested in this competition is worthwhile. Nonetheless, as a part of the FGVC workshop, our competition is more focused on the workshop itself than the Kaggle system. Let me see if I can pass this on to the workshop organizers, but please be aware that no previous FGVC workshop competitions have been involved in competition tiers.
", "votes": 1, "replies": [ { "id": 1709977, "author_name": "Faisal Alsrheed", "author_url": "", "post_date": "2022-03-02T16:49:50.193000", "content": "Thank you for the clarification.
", "votes": 1, "replies": [] }, { "id": 1735114, "author_name": "xhlulu", "author_url": "", "post_date": "2022-03-25T21:38:21.287000", "content": "Hi John. Thank you for the clarification. I know this is might be out of your control, but I just want to point out that a lot of Kagglers might not be going to be actively pursuing research or attending CVPR workshops not because they are not interested, but more because being an academic is a geographic/financial privilege.
\nWhereas a lot of folks pursuing PhDs have formal education in CS or related fields, and are from famous universities in US/EU or in prestigious national schools (e.g. IITs, PKU, SNU, etc.). Those students had the chance to work on research submitted to top conferences before even starting their PhD, and many started preparing for academia in middle school with tutoring and mentors.
\nOn the other hand, a lot of Kagglers are self-taught and rely on the free compute (free GPU and TPU from Kaggle, Colab, etc) and financial incentives (prizes from competition) to gain skills in ML. Although they might have original ideas that could benefit the research community, they might not have the guidance to work on those. Furthermore, even if they were given the chance to attend a prestigious masters abroad, they might not be able to afford it (most scholarships only cover a part of the costs, not all).
\nI think such workshops organized through Kaggle could be a good way of introducing eager students on Kaggle to academic research from throughout the world. It's easy to think that a lot of Kagglers are ML practitioners that are more interested in engineering than graduate school or academia. The truth is, however, that 43% of those who answered the 2019 survey intend to, or have attained a Master's degree, which is a research-based program in most parts of the world.
\nCompared to CVPR, Kaggle has a much more diverse user base because of the ease of access. In the 2019 survey, Parul showed in her notebook over 150 women from various African countries participating in the survey (63 from Nigeria, 29 from Kenya, 27 from Egypt) compared to 624 women in USA (and 99 in Canada). If you look at the same stats at CVPR19, only 0.2% of the submissions come from any African country, and there's only 1 attendee from Egypt, and 0 from Kenya or Nigeria. Similar observations can be made with the fact that India and Brazil are underrepresented in the CVPR survey vs Kaggle, etc.
\nI know bringing those changes would require herculean efforts. I know CVPR is working hard to improve access to underrepresented communities and those with financial barriers. I just hope that this starts a dialogue on how annual competitions by CVPR could be beneficial to more than just researchers already attending CVPR but also the broader data science community, many of whom are definitely interested in research but simply do not have the capability or incentive to participate.
", "votes": 3, "replies": [] }, { "id": 1735274, "author_name": "Sanyam Bhutani", "author_url": "", "post_date": "2022-03-26T04:21:04.830000", "content": "On a slightly silly note:
\nMaybe I can help create an unofficial Kaggle x Chai conference where we host most interesting comp winners and solutions. A virtual conference but for Kagglers where we can celebrate the awesome knowledge and Kagglers in the community. 🤔
\nThe podcast is an attempt at the same but its a broadcasted medium and not a discussion format.
", "votes": 1, "replies": [] }, { "id": 1735970, "author_name": "xhlulu", "author_url": "", "post_date": "2022-03-26T18:54:08.500000", "content": "@init27 I don't think this is a silly idea; it's great in fact. I think this sort of exposure could definitely encourage people to participate in otherwise no-prize/points competitions.
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "1707554": "I would like to thank the competition organizer and Kaggle team for hosting this competition.\n\nAs you can see from the leaderboard, I think it is a good idea to add ranking points and count towards tiers to increase the activity in this competition.\n\n\n", "1709946": "@faisalalsrheed, thanks for sharing your thoughts. We understand that ranking points would motivate more kagglers to the competition. We also want to make sure that your time invested in this competition is worthwhile. Nonetheless, as a part of the FGVC workshop, our competition is more focused on the workshop itself than the Kaggle system. Let me see if I can pass this on to the workshop organizers, but please be aware that no previous FGVC workshop competitions have been involved in competition tiers. " } }