![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1518097%2F41e74f0e4a74db406608799f9432c0d6%2Fold.jpeg?generation=1741910150585988&alt=media\")
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1518097%2F75ca43faea95f2bf8270adfdb67c0eb5%2Fnew.png?generation=1741910160614693&alt=media\")
\nBest,
\nLukas
Hello Lucas, it's been a long time, I joined the contest the first time I got your email inviting us to the contest. I appreciate your dedication to this competition, what I would like to know is what is the difference between this year's competition and last year's competition in terms of the dataset, as well as the setup of this year's competition in terms of prizes and papers to be included in the competition, thank you!
", "messages": [ { "id": "3149191", "postDate": "03/13/2025 23:14:05", "content": "Hello Lucas, it's been a long time, I joined the contest the first time I got your email inviting us to the contest. I appreciate your dedication to this competition, what I would like to know is what is the difference between this year's competition and last year's competition in terms of the dataset, as well as the setup of this year's competition in terms of prizes and papers to be included in the competition, thank you!
", "rawMarkdown": "Hello Lucas, it's been a long time, I joined the contest the first time I got your email inviting us to the contest. I appreciate your dedication to this competition, what I would like to know is what is the difference between this year's competition and last year's competition in terms of the dataset, as well as the setup of this year's competition in terms of prizes and papers to be included in the competition, thank you!", "votes": null }, { "id": "3149211", "postDate": "03/13/2025 23:57:23", "content": "Dear @liuhaixu,
\nThe training data are more or less the same, and most changes were made to the \"test\" set. The biggest differences in this year are:
PS: We have found a bug in \"source\" rasters, which caused some Sentinel2 images, including some weird artefacts (see the Image bello). We will most likely replace those within a week.
\n
\nBest,
\nLukas
Thank for you reply!
", "rawMarkdown": "Thank for you reply!", "votes": null }, { "id": "3205317", "postDate": "05/19/2025 16:36:21", "content": "Thanks for question and to host thank for answer!
", "rawMarkdown": "Thanks for question and to host thank for answer!", "votes": null } ], "comments": [ { "id": 3149211, "author_name": "picekl", "author_url": "", "post_date": "03/13/2025 23:57:23", "content": "Dear @liuhaixu,
\nThe training data are more or less the same, and most changes were made to the \"test\" set. The biggest differences in this year are:
PS: We have found a bug in \"source\" rasters, which caused some Sentinel2 images, including some weird artefacts (see the Image bello). We will most likely replace those within a week.
\n
\nBest,
\nLukas
Thank for you reply!
", "votes": null, "replies": [] } ] }, { "id": 3205317, "author_name": "", "author_url": "", "post_date": "05/19/2025 16:36:21", "content": "Thanks for question and to host thank for answer!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "3149191": "Hello Lucas, it's been a long time, I joined the contest the first time I got your email inviting us to the contest. I appreciate your dedication to this competition, what I would like to know is what is the difference between this year's competition and last year's competition in terms of the dataset, as well as the setup of this year's competition in terms of prizes and papers to be included in the competition, thank you!", "3149211": "Dear @liuhaixu,\nThe training data are more or less the same, and most changes were made to the \"test\" set. The biggest differences in this year are:\n1. There are an additional 10k plots with a considerable \"location\" shift in the test set. Therefore, there is a considerable drop in performance. We hope this will lead people to use some PO data.\n2. Human footprint values were replaced. \n\nPS: We have found a bug in \"source\" rasters, which caused some Sentinel2 images, including some weird artefacts (see the Image bello). We will most likely replace those within a week.\n \n\nBest,\nLukas", "3150103": "Thank for you reply!", "3205317": "Thanks for question and to host thank for answer!" }, "source": "meta" }