Earth-o1 learns continuous atmospheric dynamics from ungridded observations and matches operational IFS forecast skill in hindcasts.
In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Earth-o1: A Grid-free Observation-native Atmospheric World Model
Earth-o1 learns continuous atmospheric dynamics from ungridded observations and matches operational IFS forecast skill in hindcasts.