Is it possible to generate additional training datasets using generative models such as GAN, and then add them to the training using model pseudo-labeling, with no manual labeling in between, as allowed by the rules of the competition? Because I think this essentially amounts to a form of data augmentation.
", "messages": [ { "id": "3183669", "postDate": "04/21/2025 06:07:33", "content": "Is it possible to generate additional training datasets using generative models such as GAN, and then add them to the training using model pseudo-labeling, with no manual labeling in between, as allowed by the rules of the competition? Because I think this essentially amounts to a form of data augmentation.
", "rawMarkdown": "Is it possible to generate additional training datasets using generative models such as GAN, and then add them to the training using model pseudo-labeling, with no manual labeling in between, as allowed by the rules of the competition? Because I think this essentially amounts to a form of data augmentation.", "votes": null }, { "id": "3183677", "postDate": "04/21/2025 06:25:08", "content": "What in the rules makes you think this would not be allowed?
", "rawMarkdown": "What in the rules makes you think this would not be allowed?", "votes": null }, { "id": "3183679", "postDate": "04/21/2025 06:32:10", "content": "“6. EXTERNAL DATA AND TOOLS
\na. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is either publicly available and equally accessible to use by all Participants of the Competition for purposes of the competition at no cost to the other Participants, or satisfies the Reasonableness criteria as outlined in Section 2.6.b below. The ability to use External Data under this Section does not limit your other obligations under these Competition Rules, including but not limited to Section 2.8 (Winners Obligations).”
\n“Freely & publicly available external data is allowed,”
\nI think the key point is whether the private generated data is “External Data” as defined above?
\nAnd whether it needs to be made public？
Well, all your processed data is not considered external or else you would need to share the data after you preprocessed, normalized it, etc.
", "rawMarkdown": "Well, all your processed data is not considered external or else you would need to share the data after you preprocessed, normalized it, etc.", "votes": null }, { "id": "3183689", "postDate": "04/21/2025 07:00:54", "content": "Thank you very much!🤗
", "rawMarkdown": "Thank you very much!🤗", "votes": null } ], "comments": [ { "id": 3183677, "author_name": "jeroencottaar", "author_url": "", "post_date": "04/21/2025 06:25:08", "content": "What in the rules makes you think this would not be allowed?
", "votes": null, "replies": [ { "id": 3183679, "author_name": "yyyy0201", "author_url": "", "post_date": "04/21/2025 06:32:10", "content": "“6. EXTERNAL DATA AND TOOLS
\na. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is either publicly available and equally accessible to use by all Participants of the Competition for purposes of the competition at no cost to the other Participants, or satisfies the Reasonableness criteria as outlined in Section 2.6.b below. The ability to use External Data under this Section does not limit your other obligations under these Competition Rules, including but not limited to Section 2.8 (Winners Obligations).”
\n“Freely & publicly available external data is allowed,”
\nI think the key point is whether the private generated data is “External Data” as defined above?
\nAnd whether it needs to be made public？
Well, all your processed data is not considered external or else you would need to share the data after you preprocessed, normalized it, etc.
", "votes": null, "replies": [ { "id": 3183689, "author_name": "yyyy0201", "author_url": "", "post_date": "04/21/2025 07:00:54", "content": "Thank you very much!🤗
", "votes": null, "replies": [] } ] } ] } ] } ], "raw_markdown_by_id": { "3183669": "Is it possible to generate additional training datasets using generative models such as GAN, and then add them to the training using model pseudo-labeling, with no manual labeling in between, as allowed by the rules of the competition? Because I think this essentially amounts to a form of data augmentation.", "3183677": "What in the rules makes you think this would not be allowed?", "3183679": "“6. EXTERNAL DATA AND TOOLS\na. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is either publicly available and equally accessible to use by all Participants of the Competition for purposes of the competition at no cost to the other Participants, or satisfies the Reasonableness criteria as outlined in Section 2.6.b below. The ability to use External Data under this Section does not limit your other obligations under these Competition Rules, including but not limited to Section 2.8 (Winners Obligations).”\n“Freely & publicly available external data is allowed,”\nI think the key point is whether the private generated data is “External Data” as defined above?\nAnd whether it needs to be made public？", "3183686": "Well, all your processed data is not considered external or else you would need to share the data after you preprocessed, normalized it, etc.", "3183689": "Thank you very much!🤗" }, "source": "meta" }