analiza aiqbench
This research introduces STORM, a generative AI framework that reformulates data assimilation as diffusion-based Bayesian sampling to overcome the computational limits of traditional Earth system prediction methods. By employing a global-attention mechanism, the model achieves linear complexity scaling, allowing it to run massive ensembles on exascale hardware like Frontier for high-resolution climate modeling and hurricane tracking.
Global Attention with Linear Complexity for Exascale Generative Data Assimilation in Earth System Prediction
arXiv:2604.16590v2 Announce Type: replace Abstract: Accurate Earth system prediction requires state inference from incomplete observations, but conventional two-stage data assimilation (DA) is computationally prohibitive because repeated PDE-based ensemble forecasts, observation updates, and intermediate data movement limit ensemble size at high resolution. We introduce STORM, a one-stage generative AI framework that reformulates DA as diffusion-based Bayesian posterior sampling, replacing online PDE ensemble forecasts with scalable AI inference. It further combines a spatiotemporal transformer with a global-attention algorithm that reduces complexity from quadratic to linear through scalable gradient propagation, enabling high-resolution, long-context Earth modeling. STORM scales to 74,400 GPUs on Frontier with 96--99\% strong-scaling efficiency and up to 6 ExaFLOPs sustained BF16 throughput, while enabling 32,768-member ensembles for uncertainty quantification in 34 seconds on 4,096 GPUs. It scales to 20 billion spatiotemporal tokens and 177,000 temporal frames. Hurricane tracking and long-term climate reanalysis demonstrate improved accuracy, including benefits from longer temporal context and recovery of temperature extremes missed by forecast-only predictions.
Authors:Xiao Wang, Zezhong Zhang, Isaac Lyngaas, Hong-Jun Yoon, Jong-Youl Choi, Siming Liang, Janet Wang, Hristo G. Chipilski, Ashwin M. Aji, Feng Bao, Peter Jan van Leeuwen, Dan Lu, Guannan Zhang
Abstract:Accurate Earth system prediction requires state inference from incomplete observations, but conventional two-stage data assimilation (DA) is computationally prohibitive because repeated PDE-based ensemble forecasts, observation updates, and intermediate data movement limit ensemble size at high resolution. We introduce STORM, a one-stage generative AI framework that reformulates DA as diffusion-based Bayesian posterior sampling, replacing online PDE ensemble forecasts with scalable AI inference. It further combines a spatiotemporal transformer with a global-attention algorithm that reduces complexity from quadratic to linear through scalable gradient propagation, enabling high-resolution, long-context Earth modeling. STORM scales to 74,400 GPUs on Frontier with 96--99\% strong-scaling efficiency and up to 6 ExaFLOPs sustained BF16 throughput, while enabling 32,768-member ensembles for uncertainty quantification in 34 seconds on 4,096 GPUs. It scales to 20 billion spatiotemporal tokens and 177,000 temporal frames. Hurricane tracking and long-term climate reanalysis demonstrate improved accuracy, including benefits from longer temporal context and recovery of temperature extremes missed by forecast-only predictions.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2604.16590 [cs.LG]
(or arXiv:2604.16590v2 [cs.LG] for this version)
Text integral sub licența deschisă a publicației · 198 cuvinte
Subiecte conexe
Scaling Biomolecular Modeling Using Context Parallelism in NVIDIA BioNeMo
NVIDIA · 28 apr. 2026
ORBIT-2 based Weather and Climate Downscaling and Downscaled Global Forecasts on AMD Instinct
AMD · 8 iun. 2026
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · 8 oct. 2026
Triton-Based Optimization of Video Sparse Attention on ROCm
AMD · 13 iul. 2026