![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F9280694%2F9ca906aa89888c4fd40642283b18d4ba%2FScreenshot%20from%202023-11-25%2018-59-54.png?generation=1700960438000229&alt=media\")
Are publicly shared csv files (output of notebooks) allowed to be used to win?
\nHow would one reproduce the results once the competition has concluded?
\nAre these notebooks and csv files just to mislead?
\nIf you have any thoughts on this, please shared!!
", "messages": [ { "id": "2537658", "postDate": "11/25/2023 11:46:58", "content": "Are publicly shared csv files (output of notebooks) allowed to be used to win?
\nHow would one reproduce the results once the competition has concluded?
\nAre these notebooks and csv files just to mislead?
\nIf you have any thoughts on this, please shared!!
", "rawMarkdown": "Are publicly shared csv files (output of notebooks) allowed to be used to win?\n\nHow would one reproduce the results once the competition has concluded?\n\nAre these notebooks and csv files just to mislead?\n\nIf you have any thoughts on this, please shared!!", "votes": null }, { "id": "2537670", "postDate": "11/25/2023 11:56:49", "content": "Hello! It's not uncommon for participants to share their work close to the competition deadline. While some may be motivated by recognition and medals, others genuinely aim to contribute to the community. Regarding publicly shared CSV files, they might not always be easily reproducible, and top-ranking kernels can sometimes show signs of overfitting.
\nIt's essential to approach these shared resources with a critical mindset. Instead of assuming misleading intent, consider them as valuable learning opportunities. As you delve into reproducing results, focus on understanding the methodologies and adapting them to your own approach. The shared materials can serve as a springboard for your own ideas and innovations.
\nEmbrace the collaborative spirit of the community, and enjoy the learning journey! If you have any questions about specific shared content, feel free to ask for clarification or guidance. Happy exploring!
", "rawMarkdown": "Hello! It's not uncommon for participants to share their work close to the competition deadline. While some may be motivated by recognition and medals, others genuinely aim to contribute to the community. Regarding publicly shared CSV files, they might not always be easily reproducible, and top-ranking kernels can sometimes show signs of overfitting.\n\nIt's essential to approach these shared resources with a critical mindset. Instead of assuming misleading intent, consider them as valuable learning opportunities. As you delve into reproducing results, focus on understanding the methodologies and adapting them to your own approach. The shared materials can serve as a springboard for your own ideas and innovations.\n\nEmbrace the collaborative spirit of the community, and enjoy the learning journey! If you have any questions about specific shared content, feel free to ask for clarification or guidance. Happy exploring!", "votes": null }, { "id": "2537693", "postDate": "11/25/2023 12:15:57", "content": "Thanks for the answer. I really appreciate
", "rawMarkdown": "Thanks for the answer. I really appreciate", "votes": null }, { "id": "2538156", "postDate": "11/25/2023 20:23:49", "content": "My second question is whether ''links to such csv files´´ can be accepted for reproducible results in the final rankings. Or do winners have to provide a model than can be trained to produce the csv file, not just pd.read_csv(…)
", "rawMarkdown": "My second question is whether ''links to such csv files´´ can be accepted for reproducible results in the final rankings. Or do winners have to provide a model than can be trained to produce the csv file, not just pd.read_csv(...)", "votes": null }, { "id": "2538277", "postDate": "11/26/2023 01:03:37", "content": "I guess they don't see the blue banner on the notebook page asking that it not be done in the last week (see screenshot below). I wish Kaggle would enforce this. I'm in favor of not allowing any notebooks be posted the last week.
\n
agree.Instead of modifying the csv file,focus on model would be better
", "rawMarkdown": "agree.Instead of modifying the csv file,focus on model would be better", "votes": null }, { "id": "2538407", "postDate": "11/26/2023 04:43:40", "content": "I wish Kaggle would enforce this.
\nActually, I think Kaggle does it. Perhaps because the notebook was made public three days ago, there is no way to prevent subsequent versions with higher scores.😂
If someone is Prize Contender but cannot provide the detail approach, he may not get the money. However, I think he can still get the gold medal.
You're correct @lizhecheng. Kaggle doesn't allow to make a new competition notebook public, but you can push as many new versions as you wish to already shared kernel.
", "rawMarkdown": "You're correct @lizhecheng. Kaggle doesn't allow to make a new competition notebook public, but you can push as many new versions as you wish to already shared kernel.", "votes": null } ], "comments": [ { "id": 2537670, "author_name": "asimandia", "author_url": "", "post_date": "11/25/2023 11:56:49", "content": "Hello! It's not uncommon for participants to share their work close to the competition deadline. While some may be motivated by recognition and medals, others genuinely aim to contribute to the community. Regarding publicly shared CSV files, they might not always be easily reproducible, and top-ranking kernels can sometimes show signs of overfitting.
\nIt's essential to approach these shared resources with a critical mindset. Instead of assuming misleading intent, consider them as valuable learning opportunities. As you delve into reproducing results, focus on understanding the methodologies and adapting them to your own approach. The shared materials can serve as a springboard for your own ideas and innovations.
\nEmbrace the collaborative spirit of the community, and enjoy the learning journey! If you have any questions about specific shared content, feel free to ask for clarification or guidance. Happy exploring!
", "votes": null, "replies": [ { "id": 2537693, "author_name": "jeannkouagou", "author_url": "", "post_date": "11/25/2023 12:15:57", "content": "Thanks for the answer. I really appreciate
", "votes": null, "replies": [] }, { "id": 2538156, "author_name": "jeannkouagou", "author_url": "", "post_date": "11/25/2023 20:23:49", "content": "My second question is whether ''links to such csv files´´ can be accepted for reproducible results in the final rankings. Or do winners have to provide a model than can be trained to produce the csv file, not just pd.read_csv(…)
", "votes": null, "replies": [ { "id": 2538295, "author_name": "blumenkranz7", "author_url": "", "post_date": "11/26/2023 01:54:36", "content": "agree.Instead of modifying the csv file,focus on model would be better
", "votes": null, "replies": [] }, { "id": 2538408, "author_name": "lizhecheng", "author_url": "", "post_date": "11/26/2023 04:46:27", "content": "If someone is Prize Contender but cannot provide the detail approach, he may not get the money. However, I think he can still get the gold medal.
I guess they don't see the blue banner on the notebook page asking that it not be done in the last week (see screenshot below). I wish Kaggle would enforce this. I'm in favor of not allowing any notebooks be posted the last week.
\n
I wish Kaggle would enforce this.
\nActually, I think Kaggle does it. Perhaps because the notebook was made public three days ago, there is no way to prevent subsequent versions with higher scores.😂
You're correct @lizhecheng. Kaggle doesn't allow to make a new competition notebook public, but you can push as many new versions as you wish to already shared kernel.
", "votes": null, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2537658": "Are publicly shared csv files (output of notebooks) allowed to be used to win?\n\nHow would one reproduce the results once the competition has concluded?\n\nAre these notebooks and csv files just to mislead?\n\nIf you have any thoughts on this, please shared!!", "2537670": "Hello! It's not uncommon for participants to share their work close to the competition deadline. While some may be motivated by recognition and medals, others genuinely aim to contribute to the community. Regarding publicly shared CSV files, they might not always be easily reproducible, and top-ranking kernels can sometimes show signs of overfitting.\n\nIt's essential to approach these shared resources with a critical mindset. Instead of assuming misleading intent, consider them as valuable learning opportunities. As you delve into reproducing results, focus on understanding the methodologies and adapting them to your own approach. The shared materials can serve as a springboard for your own ideas and innovations.\n\nEmbrace the collaborative spirit of the community, and enjoy the learning journey! If you have any questions about specific shared content, feel free to ask for clarification or guidance. Happy exploring!", "2537693": "Thanks for the answer. I really appreciate", "2538156": "My second question is whether ''links to such csv files´´ can be accepted for reproducible results in the final rankings. Or do winners have to provide a model than can be trained to produce the csv file, not just pd.read_csv(...)", "2538277": "I guess they don't see the blue banner on the notebook page asking that it not be done in the last week (see screenshot below). I wish Kaggle would enforce this. I'm in favor of not allowing any notebooks be posted the last week.\n\n", "2538295": "agree.Instead of modifying the csv file,focus on model would be better", "2538407": "```I wish Kaggle would enforce this.```\nActually, I think Kaggle does it. Perhaps because the notebook was made public three days ago, there is no way to prevent subsequent versions with higher scores.😂", "2538408": "If someone is ``Prize Contender`` but cannot provide the detail approach, he may not get the money. However, I think he can still get the gold medal.", "2538598": "You're correct @lizhecheng. Kaggle doesn't allow to make a new competition notebook public, but you can push as many new versions as you wish to already shared kernel." }, "source": "meta" }