Understanding machine learning weather prediction by designing a cost-efficient model with knowledge-oriented modules
Minjong Cheon, Jeong-Hwan Kim, Yumi Choi, Yo-Hwan Choi, Seon-Yu Kang, Jeong-Gil Lee, Yoo-Geun Ham, Jin Young Kim, Daehyun Kang · 2025
Read Understanding machine learning weather prediction by designing a cost-efficient model with knowledge-oriented modules by Minjong Cheon, Jeong-Hwan Kim, Yumi Choi, Yo-Hwan Choi, Seon-Yu Kang, Jeong-Gil Lee, Yoo-Geun Ham, Jin Young Kim, Daehyun Kang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-12819519.txt.
Description
Minjong Cheon and co-authors build a cheap AI weather model and use it to learn how such models work. The study appeared in 2025. Top AI weather models need huge compute to train, which blocks deep tests of their parts. Their skill also varies by place and by trait in ways no one had mapped.
The new model pairs two add-on parts with a known image net core. It matched top AI models and beat a top physics model out to ten days, at far lower train cost. Tests showed one add-on helps winds move across the map. The other helps capture rising air and storms.
Catalog record
| CREATOR | Minjong Cheon, Jeong-Hwan Kim, Yumi Choi, Yo-Hwan Choi, Seon-Yu Kang, Jeong-Gil Lee, Yoo-Geun Ham, Jin Young Kim, Daehyun Kang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |