analiza aiqbench
This research advances neural operator architectures by integrating spherical geometry and physics-informed constraints to solve the specific challenges of global-to-regional meteorological downscaling. By embedding differentiable atmospheric laws and utilizing Laplacian decomposition, the model achieves superior accuracy and zero-shot generalization for climate forecasting without relying on traditional numerical weather prediction models. The work represents a significant step in applying machine learning to complex spatiotemporal mappings where geometric fidelity and physical consistency are paramount.
Spherical Physics-informed Neural Operator with Multi-scale Coupling for Meteorological Downscaling
arXiv:2610.24708v1 Announce Type: cross Abstract: Meteorological downscaling is crucial for high-resolution regional climate forecasting and disaster early warning. While neural operators have emerged as a promising paradigm for modeling complex spatiotemporal mappings, existing frameworks often struggle with spherical manifold geometric distortions, inherent atmospheric multi-scale coupling mismatches, and lack of explicit atmospheric laws. We propose the Spherical Physics-informed Neural Operator, which utilizes a Spherical Laplacian Decomposition to partition atmospheric fields into hierarchical frequency components, maintaining exact point-wise correspondence across scales. To evaluate these representations at arbitrary locations, we introduce a localized spherical integral operator that approximates continuous kernel transforms via geometry-aware attention. Dynamical consistency is further enforced by embedding differentiable constraints into the learning process. Extensive experiments demonstrate that our framework attains superior accuracy and zero-shot generalization across various meteorological variables and unseen queries, representing a robust and interpretable solution for global-to-regional meteorological downscaling.
Abstract:Meteorological downscaling is crucial for high-resolution regional climate forecasting and disaster early warning. While neural operators have emerged as a promising paradigm for modeling complex spatiotemporal mappings, existing frameworks often struggle with spherical manifold geometric distortions, inherent atmospheric multi-scale coupling mismatches, and lack of explicit atmospheric laws. We propose the Spherical Physics-informed Neural Operator, which utilizes a Spherical Laplacian Decomposition to partition atmospheric fields into hierarchical frequency components, maintaining exact point-wise correspondence across scales. To evaluate these representations at arbitrary locations, we introduce a localized spherical integral operator that approximates continuous kernel transforms via geometry-aware attention. Dynamical consistency is further enforced by embedding differentiable constraints into the learning process. Extensive experiments demonstrate that our framework attains superior accuracy and zero-shot generalization across various meteorological variables and unseen queries, representing a robust and interpretable solution for global-to-regional meteorological downscaling.
Published in IJCAI 2026 (AI4Tech: AI Enabling Technologies). 9 pages, 4 figures
Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph)
arXiv:2610.24708 [cs.AI]
(or arXiv:2610.24708v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.24708 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence(IJCAI-26), pp. 6600-6608, 2026
https://doi.org/10.24963/ijcai.2026/734 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 275 cuvinte
Subiecte conexe
News
Quantum Computing Report · 25 iul. 2026
HAMNO: A Hierarchical Adaptive Multi-scale Neural Operator with Physics-Informed Learning for Dynamical Systems
arXiv · 12 iun. 2026
Does preservation make sense before we know how to revive?
LessWrong · 15 iun. 2026
Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk
Latent Space · 8 oct. 2026