STORM achieves linear-complexity global attention in a generative DA framework, scaling to 20 billion tokens and 1.6 ExaFLOPs on 32k GPUs for km-scale Earth modeling.
A generalization of Lorenz’s model for the predictability of flows with many scales of motion
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Global Attention with Linear Complexity for Exascale Generative Data Assimilation in Earth System Prediction
STORM achieves linear-complexity global attention in a generative DA framework, scaling to 20 billion tokens and 1.6 ExaFLOPs on 32k GPUs for km-scale Earth modeling.