Explore a groundbreaking conference talk on AI-powered global weather forecasting presented at ICBS2024. Delve into the innovative Pangu-Weather system, which utilizes 3D deep neural networks to achieve accurate medium-range predictions. Learn how this AI-based method outperforms traditional numerical weather prediction (NWP) systems, including the world-renowned ECMWF operational integrated forecasting system. Discover the application of Earth-specific priors and hierarchical temporal aggregation strategies that enable Pangu-Weather to handle complex weather patterns and reduce accumulation errors. Gain insights into the system's effectiveness in extreme weather forecasting, ensemble predictions, and tropical cyclone tracking. Understand the potential of this technology to revolutionize weather forecasting, offering significant improvements in accuracy and computational efficiency compared to current methods.
Accurate Medium-Range Global Weather Forecasting with 3D Deep Neural Networks
Overview
Syllabus
Lingxi Xie: Accurate Medium-Range Global Weather Forecasting with 3D Deep Neural Networks #ICBS2024
Taught by
BIMSA
Related Courses
-
Pangu-Weather: Accurate Medium-Range Global Weather Forecasting with 3D Deep Neural Networks
-
Weather Forecasting Using Deep Learning - A Paradigm Shift with MetNet-3
-
Axial Attention & MetNet - A Neural Weather Model for Precipitation Forecasting
-
How AI is Changing Weather Forecasting Using Ocean Data
-
Forecasting Weather with Neural Prophet and Python
-
AI and the Second Revolution in Weather and Climate Prediction
