{ "id": 603421, "title": "shake up or not?", "url": "/competitions/cmi-detect-behavior-with-sensor-data/discussion/603421", "author_name": "", "post_date": "2025-09-02T12:39:18.059792600Z", "votes": 3, "comment_count": 7, "views": 0, "content": "

Personally I think there will not be a significant change in the leaderboard, what about your opinions?

", "messages": [ { "id": "3280292", "postDate": "09/02/2025 12:39:18", "content": "

Personally I think there will not be a significant change in the leaderboard, what about your opinions?

", "rawMarkdown": "Personally I think there will not be a significant change in the leaderboard, what about your opinions?", "votes": null }, { "id": "3280327", "postDate": "09/02/2025 13:20:25", "content": "

I think we will have a shakeup to a small extent - top teams will be safe @belalemadhussein

", "rawMarkdown": "I think we will have a shakeup to a small extent - top teams will be safe @belalemadhussein", "votes": null }, { "id": "3280333", "postDate": "09/02/2025 13:26:29", "content": "

yes, maybe except for who have done post-processing that leads to overfitting

", "rawMarkdown": "yes, maybe except for who have done post-processing that leads to overfitting", "votes": null }, { "id": "3280577", "postDate": "09/02/2025 21:24:53", "content": "

Short answer, not that much i think 😞

\n

from previous history on CMI - i have overfitted to lb, but we all know the difference between folds and changing seeds led to different results, and huge shakeup is coming then. Problematic Internet USE

", "rawMarkdown": "Short answer, not that much i think 😞\n\nfrom previous history on CMI - i have overfitted to lb, but we all know the difference between folds and changing seeds led to different results, and huge shakeup is coming then. [Problematic Internet USE](https://www.kaggle.com/competitions/child-mind-institute-problematic-internet-use/leaderboard?tab=public)", "votes": null }, { "id": "3280588", "postDate": "09/02/2025 22:07:17", "content": "

In the last CMI, I shakedown from 2nd place to 1075th. How I felt at that time:
\n\"OMG!!! Aaaaahhhhhhhhhh!\"

\n

P.S. This competition required me to do a lot of original implementation, such as model structure and data aug, but I gained a lot of experience and improved my skills, so I'm satisfied now.
\n(This is just how I felt before my private score was announced…)

\n

\"\"

", "rawMarkdown": "In the last CMI, I shakedown from 2nd place to 1075th. How I felt at that time:\n\"OMG!!! Aaaaahhhhhhhhhh!\"\n\nP.S. This competition required me to do a lot of original implementation, such as model structure and data aug, but I gained a lot of experience and improved my skills, so I'm satisfied now.\n(This is just how I felt before my private score was announced...)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3802145%2F022c91947a9e56ea8e87154bc9a2781c%2F2025-09-03%20064952.jpg?generation=1756849923055900&alt=media)", "votes": null }, { "id": "3280590", "postDate": "09/02/2025 22:16:48", "content": "

it was one of the big shakeup \"For me at least\"
\ndown from 1 to 1207,
\nbut i was expecting that, each param change lead to completely different results

", "rawMarkdown": "it was one of the big shakeup \"For me at least\" \ndown from 1 to 1207,\nbut i was expecting that, each param change lead to completely different results", "votes": null }, { "id": "3280657", "postDate": "09/03/2025 02:09:11", "content": "

@hideyukizushi always thank you for providing good public resources!

", "rawMarkdown": "hideyukizushi always thank you for providing good public resources!", "votes": null }, { "id": "3280705", "postDate": "09/03/2025 03:41:22", "content": "

Hi! @tom99763
\nthank you for respecting my code.

\n

I'm always grateful to those who share their baselines with me.
\nI've learned a lot from wonderful code, so I hope the culture of sharing ideas on Kaggle as concrete code will continue in the future.

", "rawMarkdown": "Hi! @tom99763 \nthank you for respecting my code. \n\nI'm always grateful to those who share their baselines with me.\nI've learned a lot from wonderful code, so I hope the culture of sharing ideas on Kaggle as concrete code will continue in the future.", "votes": null } ], "comments": [ { "id": 3280327, "author_name": "ravi20076", "author_url": "", "post_date": "09/02/2025 13:20:25", "content": "

I think we will have a shakeup to a small extent - top teams will be safe @belalemadhussein

", "votes": null, "replies": [ { "id": 3280333, "author_name": "belalemadhussein", "author_url": "", "post_date": "09/02/2025 13:26:29", "content": "

yes, maybe except for who have done post-processing that leads to overfitting

", "votes": null, "replies": [] } ] }, { "id": 3280577, "author_name": "letemoin", "author_url": "", "post_date": "09/02/2025 21:24:53", "content": "

Short answer, not that much i think 😞

\n

from previous history on CMI - i have overfitted to lb, but we all know the difference between folds and changing seeds led to different results, and huge shakeup is coming then. Problematic Internet USE

", "votes": null, "replies": [ { "id": 3280588, "author_name": "hideyukizushi", "author_url": "", "post_date": "09/02/2025 22:07:17", "content": "

In the last CMI, I shakedown from 2nd place to 1075th. How I felt at that time:
\n\"OMG!!! Aaaaahhhhhhhhhh!\"

\n

P.S. This competition required me to do a lot of original implementation, such as model structure and data aug, but I gained a lot of experience and improved my skills, so I'm satisfied now.
\n(This is just how I felt before my private score was announced…)

\n

\"\"

", "votes": null, "replies": [ { "id": 3280590, "author_name": "letemoin", "author_url": "", "post_date": "09/02/2025 22:16:48", "content": "

it was one of the big shakeup \"For me at least\"
\ndown from 1 to 1207,
\nbut i was expecting that, each param change lead to completely different results

", "votes": null, "replies": [] }, { "id": 3280657, "author_name": "tom99763", "author_url": "", "post_date": "09/03/2025 02:09:11", "content": "

@hideyukizushi always thank you for providing good public resources!

", "votes": null, "replies": [ { "id": 3280705, "author_name": "hideyukizushi", "author_url": "", "post_date": "09/03/2025 03:41:22", "content": "

Hi! @tom99763
\nthank you for respecting my code.

\n

I'm always grateful to those who share their baselines with me.
\nI've learned a lot from wonderful code, so I hope the culture of sharing ideas on Kaggle as concrete code will continue in the future.

", "votes": null, "replies": [] } ] } ] } ] } ], "raw_markdown_by_id": { "3280292": "Personally I think there will not be a significant change in the leaderboard, what about your opinions?", "3280327": "I think we will have a shakeup to a small extent - top teams will be safe @belalemadhussein", "3280333": "yes, maybe except for who have done post-processing that leads to overfitting", "3280577": "Short answer, not that much i think 😞\n\nfrom previous history on CMI - i have overfitted to lb, but we all know the difference between folds and changing seeds led to different results, and huge shakeup is coming then. [Problematic Internet USE](https://www.kaggle.com/competitions/child-mind-institute-problematic-internet-use/leaderboard?tab=public)", "3280588": "In the last CMI, I shakedown from 2nd place to 1075th. How I felt at that time:\n\"OMG!!! Aaaaahhhhhhhhhh!\"\n\nP.S. This competition required me to do a lot of original implementation, such as model structure and data aug, but I gained a lot of experience and improved my skills, so I'm satisfied now.\n(This is just how I felt before my private score was announced...)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3802145%2F022c91947a9e56ea8e87154bc9a2781c%2F2025-09-03%20064952.jpg?generation=1756849923055900&alt=media)", "3280590": "it was one of the big shakeup \"For me at least\" \ndown from 1 to 1207,\nbut i was expecting that, each param change lead to completely different results", "3280657": "hideyukizushi always thank you for providing good public resources!", "3280705": "Hi! @tom99763 \nthank you for respecting my code. \n\nI'm always grateful to those who share their baselines with me.\nI've learned a lot from wonderful code, so I hope the culture of sharing ideas on Kaggle as concrete code will continue in the future." }, "source": "meta" }