https://github.com/shivamkumar818/cascadecas
Cascading Machine Learning for Long-Term Forecasting (ONGOING)
https://github.com/shivamkumar818/cascadecas
Last synced: about 2 months ago
JSON representation
Cascading Machine Learning for Long-Term Forecasting (ONGOING)
- Host: GitHub
- URL: https://github.com/shivamkumar818/cascadecas
- Owner: shivamkumar818
- Created: 2025-03-26T18:59:56.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2025-03-26T19:02:53.000Z (2 months ago)
- Last Synced: 2025-03-26T20:22:35.423Z (2 months ago)
- Homepage:
- Size: 3.88 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
FuXi: A Cascading Machine Learning Forecasting System for 15-Day Global Weather Prediction
Overview
FuXi is an advanced machine learning-based weather forecasting system designed to improve long-range predictions. Leveraging state-of-the-art ML models, it aims to outperform traditional physics-based forecasting methods, such as ECMWF HRES, particularly in 15-day weather forecasts.
Features
ML-Powered Forecasting: Uses deep learning models to generate accurate long-range weather predictions.
Error Mitigation: Implements autoregressive multi-timestep loss techniques to reduce accumulation errors in extended forecasts.
High Spatial Resolution: Achieves comparable or superior accuracy to ECMWF ensemble models.
Scalability: Designed for deployment in real-world meteorological applications.
Technologies Used
Deep Learning (CNN, LSTM, Transformers)
Time-Series Forecasting Techniques
Autoregressive Modeling
High-Performance Computing (HPC) for large-scale simulations