Alibaba DAMO Academy Launches Baguan, AI-Based Weather Estimated Model

Baguan's technological introduction (photo: Alibaba DAMO Academy)

JAKARTA - Alibaba DAMO Academy, a research and development institute from Alibaba Group, today announced the official launch of Baguan, an artificial intelligence (AI)-supported weather forecast model.

With existing machine learning technology, Baguan offers accuracy in weather forecasts, ranging from one hour to ten days ahead, and will be updated every hour.

Wotao Yin, Director of Decision Intelligence Lab at Alibaba DAMO Academy said that Baguan is very suitable for applications related to climate knowledge, estimates of electricity loads, estimates of renewable energy, and prevention of natural disasters.

"This advanced technology not only helps improve climate science, but also provides benefits for sustainable practices in various sectors such as renewable energy and agriculture," said Wotao.

The technical basis of Baguan is the innovative use of the structure of Siamese Masked Autoencoders (SiammaE) and the previously trained powerful methodology. This innovation allows this model to uncover the complicated patterns obtained from complex atmospheric data.

Baguan took advantage of the ERA5, a global weather atmosphere reanalysis of the European Center for Medium-Range Weather Forecast (ECMWF) since 1979 until now, to build a basic weather forecast model.

Baguan was then refined by major regional meteorological indicators such as regional temperature, irradiation, and wind speed in certain regions. This global-regional modeling approach not only improves forecast accuracy, but also adapts its insight to existing local conditions.

We have years of research experience in mathematical modeling, time-series estimation, and AI that can be described, which helps us build regional weather forecast models with high precision," Yin said.

Alibaba DAMO Academy also promised to continue to improve performance for key weather indicators such as cloud cover, extreme wind speed, and rainfall, developing new technologies for different climate scenario analysis.