Hello Kagglers,
\nWe have less than 100 participants for this competition, which could have been a lot more had Kaggle given some credit to this competition in form of medals (gold/silver/bronze). Cash prizes are not always incentive for everyone, but there should be some support for competitions like this which hope to lesson the evils in our society. Curious to know everyone's view on this.
\nThanks and Regards,
\nOld Monk
Hello Kagglers,
\nWe have less than 100 participants for this competition, which could have been a lot more had Kaggle given some credit to this competition in form of medals (gold/silver/bronze). Cash prizes are not always incentive for everyone, but there should be some support for competitions like this which hope to lesson the evils in our society. Curious to know everyone's view on this.
\nThanks and Regards,
\nOld Monk
There has already been discussion about it in this competition No medals, points, prizes for this serious issue? without any response from kaggle or competition host and similar discussion in Herbarium 2022 where at least the competition host explained why FGVC9 competitions don't offer competition tiers or rewards.
\nI agree that the participation is low (even lower than last year) and any kind of incentive (tiers, swag, money, …) would help to attract more participants and could lead to better results. It would be nice if Kaggle would somehow support competitions like this but I guess that would require also some interest from the competition host and they seem to be not interested.
", "rawMarkdown": "There has already been discussion about it in this competition [No medals, points, prizes for this serious issue?](https://www.kaggle.com/competitions/hotel-id-to-combat-human-trafficking-2022-fgvc9/discussion/314885) without any response from kaggle or competition host and similar discussion in [Herbarium 2022](https://www.kaggle.com/competitions/herbarium-2022-fgvc9/discussion/310277) where at least the competition host explained why FGVC9 competitions don't offer competition tiers or rewards.\n\nI agree that the participation is low (even lower than last year) and any kind of incentive (tiers, swag, money, ...) would help to attract more participants and could lead to better results. It would be nice if Kaggle would somehow support competitions like this but I guess that would require also some interest from the competition host and they seem to be not interested.", "votes": null }, { "id": "1805766", "postDate": "05/30/2022 13:24:30", "content": "We're definitely interested -- but unfortunately it's not within our power to get Kaggle to make this a competition with tiers/swag/money. None of the FGVC competitions have that and have not previously, I believe (I could be wrong) -- as explained in the Herbarium post, our competitions are very tied to the FGVC workshop and the recognition system is linked to the workshop. Winning competitions receive recognition via the workshop. I'm sorry that may not be aligned with the Kaggle incentive structure, but it's not something we can effect.
\nAs to the low participation in this competition relative to others -- it's something I'm actually very interested in. I'm running another competition at FGVC which is a more traditional classification problem (Sorghum-100) and it has much higher engagement despite us not doing anything different in how it's promoted. I believe that this problem (the human trafficking/hotel recognition, especially with occlusions in the query images) is generally more challenging and has more of a barrier to entry, leading to fewer submissions.
\nNonetheless, last year we were able to learn a lot from the high scoring participants that informed how we're doing hotel recognition in the real world system. We're excited to do the same this year.
", "rawMarkdown": "We're definitely interested -- but unfortunately it's not within our power to get Kaggle to make this a competition with tiers/swag/money. None of the FGVC competitions have that and have not previously, I believe (I could be wrong) -- as explained in the Herbarium post, our competitions are very tied to the FGVC workshop and the recognition system is linked to the workshop. Winning competitions receive recognition via the workshop. I'm sorry that may not be aligned with the Kaggle incentive structure, but it's not something we can effect.\n\nAs to the low participation in this competition relative to others -- it's something I'm actually very interested in. I'm running another competition at FGVC which is a more traditional classification problem ([Sorghum-100](https://www.kaggle.com/c/sorghum-id-fgvc-9/discussion/314073)) and it has *much* higher engagement despite us not doing anything different in how it's promoted. I believe that this problem (the human trafficking/hotel recognition, especially with occlusions in the query images) is generally more challenging and has more of a barrier to entry, leading to fewer submissions.\n\nNonetheless, last year we were able to learn a lot from the high scoring participants that informed how we're doing hotel recognition in the real world system. We're excited to do the same this year.", "votes": null }, { "id": "1806090", "postDate": "05/30/2022 19:20:15", "content": "Hi Old Monk,
\nwhen I made a topic about \"Underrated Competitions. Join Just for Joy (of Learning).\"
\nI included this competition and the others Fine Grained (FGVC) on the list.
Unfortunately, not many users adhere to the cause.
", "rawMarkdown": "Hi Old Monk,\nwhen I made a topic about \"Underrated Competitions. Join Just for Joy (of Learning).\"\nI included this competition and the others Fine Grained (FGVC) on the list.\n\nUnfortunately, not many users adhere to the cause.", "votes": null }, { "id": "1806248", "postDate": "05/31/2022 00:37:13", "content": "Would it help if someone could raise money? My impression was no, this is an issue of bureaucracy, not money.
", "rawMarkdown": "Would it help if someone could raise money? My impression was no, this is an issue of bureaucracy, not money.", "votes": null }, { "id": "1806676", "postDate": "05/31/2022 11:56:13", "content": "I would also be interested why it's not possible. I understand that convincing Kaggle to sponsor competition would be tricky (but you can ask, it's for a good cause they might be interested), making it tier competition might be more expensive but offering CVPR/FGVC swag should not be a problem? Or making a some swag items (t-shirts/hoodies/cups, …) specifically for this competition, print on demand is cheap so all you need is idea.
\nOr at least free pass for CVPR for the winners. That's what I find most confusing, to be able to present your winning solution you have to register (and pay) for CVPR in advance because \"PLEASE NOTE: CVPR frequently sells out early, and we cannot guarantee CVPR registration after the competition's end\" otherwise you can't attend the workshop and receive the recognition there.
\nBut of course if you are happy with the results of this competition anyway there is no need to change anything. Even though the participation is low you probably get some nice and useful solutions and that's what's important. Prizes should not matter and the motivation here (at least for me) is to help and hopefully contribute to solving a problem. But making this competition more attractive might help to get more attention, discussion and more people with better ideas.
\nThe significantly higher participation in Sorghum -100 is interesting, maybe classification problem is more approachable than image similarity? I also wonder what amount of participants attend the FGVC workshop. Maybe in the post competition survey you can ask why people attend specific competition and if they considered others from the FGVC competitions.
", "rawMarkdown": "I would also be interested why it's not possible. I understand that convincing Kaggle to sponsor competition would be tricky (but you can ask, it's for a good cause they might be interested), making it tier competition might be more expensive but offering CVPR/FGVC swag should not be a problem? Or making a some swag items (t-shirts/hoodies/cups, ...) specifically for this competition, print on demand is cheap so all you need is idea. \n\nOr at least free pass for CVPR for the winners. That's what I find most confusing, to be able to present your winning solution you have to register (and pay) for CVPR in advance because \"PLEASE NOTE: CVPR frequently sells out early, and we cannot guarantee CVPR registration after the competition's end\" otherwise you can't attend the workshop and receive the recognition there.\n\nBut of course if you are happy with the results of this competition anyway there is no need to change anything. Even though the participation is low you probably get some nice and useful solutions and that's what's important. Prizes should not matter and the motivation here (at least for me) is to help and hopefully contribute to solving a problem. But making this competition more attractive might help to get more attention, discussion and more people with better ideas.\n\nThe significantly higher participation in Sorghum -100 is interesting, maybe classification problem is more approachable than image similarity? I also wonder what amount of participants attend the FGVC workshop. Maybe in the post competition survey you can ask why people attend specific competition and if they considered others from the FGVC competitions.", "votes": null }, { "id": "1807490", "postDate": "06/01/2022 04:38:54", "content": "Agree @mpwolke !
", "rawMarkdown": "Agree @mpwolke !", "votes": null }, { "id": "1807492", "postDate": "06/01/2022 04:39:36", "content": "Great point @michaln, thanks for sharing your views here!
", "rawMarkdown": "Great point @michaln, thanks for sharing your views here!", "votes": null } ], "comments": [ { "id": 1805735, "author_name": "michaln", "author_url": "", "post_date": "05/30/2022 13:08:24", "content": "There has already been discussion about it in this competition No medals, points, prizes for this serious issue? without any response from kaggle or competition host and similar discussion in Herbarium 2022 where at least the competition host explained why FGVC9 competitions don't offer competition tiers or rewards.
\nI agree that the participation is low (even lower than last year) and any kind of incentive (tiers, swag, money, …) would help to attract more participants and could lead to better results. It would be nice if Kaggle would somehow support competitions like this but I guess that would require also some interest from the competition host and they seem to be not interested.
", "votes": null, "replies": [ { "id": 1805766, "author_name": "abbystylianou", "author_url": "", "post_date": "05/30/2022 13:24:30", "content": "We're definitely interested -- but unfortunately it's not within our power to get Kaggle to make this a competition with tiers/swag/money. None of the FGVC competitions have that and have not previously, I believe (I could be wrong) -- as explained in the Herbarium post, our competitions are very tied to the FGVC workshop and the recognition system is linked to the workshop. Winning competitions receive recognition via the workshop. I'm sorry that may not be aligned with the Kaggle incentive structure, but it's not something we can effect.
\nAs to the low participation in this competition relative to others -- it's something I'm actually very interested in. I'm running another competition at FGVC which is a more traditional classification problem (Sorghum-100) and it has much higher engagement despite us not doing anything different in how it's promoted. I believe that this problem (the human trafficking/hotel recognition, especially with occlusions in the query images) is generally more challenging and has more of a barrier to entry, leading to fewer submissions.
\nNonetheless, last year we were able to learn a lot from the high scoring participants that informed how we're doing hotel recognition in the real world system. We're excited to do the same this year.
", "votes": null, "replies": [] }, { "id": 1806248, "author_name": "dfrankow", "author_url": "", "post_date": "05/31/2022 00:37:13", "content": "Would it help if someone could raise money? My impression was no, this is an issue of bureaucracy, not money.
", "votes": null, "replies": [] }, { "id": 1806676, "author_name": "michaln", "author_url": "", "post_date": "05/31/2022 11:56:13", "content": "I would also be interested why it's not possible. I understand that convincing Kaggle to sponsor competition would be tricky (but you can ask, it's for a good cause they might be interested), making it tier competition might be more expensive but offering CVPR/FGVC swag should not be a problem? Or making a some swag items (t-shirts/hoodies/cups, …) specifically for this competition, print on demand is cheap so all you need is idea.
\nOr at least free pass for CVPR for the winners. That's what I find most confusing, to be able to present your winning solution you have to register (and pay) for CVPR in advance because \"PLEASE NOTE: CVPR frequently sells out early, and we cannot guarantee CVPR registration after the competition's end\" otherwise you can't attend the workshop and receive the recognition there.
\nBut of course if you are happy with the results of this competition anyway there is no need to change anything. Even though the participation is low you probably get some nice and useful solutions and that's what's important. Prizes should not matter and the motivation here (at least for me) is to help and hopefully contribute to solving a problem. But making this competition more attractive might help to get more attention, discussion and more people with better ideas.
\nThe significantly higher participation in Sorghum -100 is interesting, maybe classification problem is more approachable than image similarity? I also wonder what amount of participants attend the FGVC workshop. Maybe in the post competition survey you can ask why people attend specific competition and if they considered others from the FGVC competitions.
", "votes": null, "replies": [] }, { "id": 1807492, "author_name": "saurabhbagchi", "author_url": "", "post_date": "06/01/2022 04:39:36", "content": "Great point @michaln, thanks for sharing your views here!
", "votes": null, "replies": [] } ] }, { "id": 1806090, "author_name": "mpwolke", "author_url": "", "post_date": "05/30/2022 19:20:15", "content": "Hi Old Monk,
\nwhen I made a topic about \"Underrated Competitions. Join Just for Joy (of Learning).\"
\nI included this competition and the others Fine Grained (FGVC) on the list.
Unfortunately, not many users adhere to the cause.
", "votes": null, "replies": [ { "id": 1807490, "author_name": "saurabhbagchi", "author_url": "", "post_date": "06/01/2022 04:38:54", "content": "Agree @mpwolke !
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1805625": "Hello Kagglers,\n\nWe have less than 100 participants for this competition, which could have been a lot more had Kaggle given some credit to this competition in form of medals (gold/silver/bronze). Cash prizes are not always incentive for everyone, but there should be some support for competitions like this which hope to lesson the evils in our society. Curious to know everyone's view on this.\n\nThanks and Regards,\nOld Monk", "1805735": "There has already been discussion about it in this competition [No medals, points, prizes for this serious issue?](https://www.kaggle.com/competitions/hotel-id-to-combat-human-trafficking-2022-fgvc9/discussion/314885) without any response from kaggle or competition host and similar discussion in [Herbarium 2022](https://www.kaggle.com/competitions/herbarium-2022-fgvc9/discussion/310277) where at least the competition host explained why FGVC9 competitions don't offer competition tiers or rewards.\n\nI agree that the participation is low (even lower than last year) and any kind of incentive (tiers, swag, money, ...) would help to attract more participants and could lead to better results. It would be nice if Kaggle would somehow support competitions like this but I guess that would require also some interest from the competition host and they seem to be not interested.", "1805766": "We're definitely interested -- but unfortunately it's not within our power to get Kaggle to make this a competition with tiers/swag/money. None of the FGVC competitions have that and have not previously, I believe (I could be wrong) -- as explained in the Herbarium post, our competitions are very tied to the FGVC workshop and the recognition system is linked to the workshop. Winning competitions receive recognition via the workshop. I'm sorry that may not be aligned with the Kaggle incentive structure, but it's not something we can effect.\n\nAs to the low participation in this competition relative to others -- it's something I'm actually very interested in. I'm running another competition at FGVC which is a more traditional classification problem ([Sorghum-100](https://www.kaggle.com/c/sorghum-id-fgvc-9/discussion/314073)) and it has *much* higher engagement despite us not doing anything different in how it's promoted. I believe that this problem (the human trafficking/hotel recognition, especially with occlusions in the query images) is generally more challenging and has more of a barrier to entry, leading to fewer submissions.\n\nNonetheless, last year we were able to learn a lot from the high scoring participants that informed how we're doing hotel recognition in the real world system. We're excited to do the same this year.", "1806090": "Hi Old Monk,\nwhen I made a topic about \"Underrated Competitions. Join Just for Joy (of Learning).\"\nI included this competition and the others Fine Grained (FGVC) on the list.\n\nUnfortunately, not many users adhere to the cause.", "1806248": "Would it help if someone could raise money? My impression was no, this is an issue of bureaucracy, not money.", "1806676": "I would also be interested why it's not possible. I understand that convincing Kaggle to sponsor competition would be tricky (but you can ask, it's for a good cause they might be interested), making it tier competition might be more expensive but offering CVPR/FGVC swag should not be a problem? Or making a some swag items (t-shirts/hoodies/cups, ...) specifically for this competition, print on demand is cheap so all you need is idea. \n\nOr at least free pass for CVPR for the winners. That's what I find most confusing, to be able to present your winning solution you have to register (and pay) for CVPR in advance because \"PLEASE NOTE: CVPR frequently sells out early, and we cannot guarantee CVPR registration after the competition's end\" otherwise you can't attend the workshop and receive the recognition there.\n\nBut of course if you are happy with the results of this competition anyway there is no need to change anything. Even though the participation is low you probably get some nice and useful solutions and that's what's important. Prizes should not matter and the motivation here (at least for me) is to help and hopefully contribute to solving a problem. But making this competition more attractive might help to get more attention, discussion and more people with better ideas.\n\nThe significantly higher participation in Sorghum -100 is interesting, maybe classification problem is more approachable than image similarity? I also wonder what amount of participants attend the FGVC workshop. Maybe in the post competition survey you can ask why people attend specific competition and if they considered others from the FGVC competitions.", "1807490": "Agree @mpwolke !", "1807492": "Great point @michaln, thanks for sharing your views here!" }, "source": "meta" }