(Yicai Global) July 31 -- Fengwu, a machine learning model developed by the Shanghai Artificial Intelligence Laboratory, has surpassed European and American equivalents in predicting the moves of Doksuri, the strongest typhoon to hit China this year.

Between July 21 and July 27, Fengwu's forecast of the tropical cyclone's path was just 38.7 kilometers wrong on average, whereas the corresponding numbers were 54.1 km for a model of the European Center for Medium-Range Weather Forecasts and 55 km for that of the United States National Centers for Environmental Prediction, Yicai Global learned from the lab in the eastern city.

Doksuri, the fifth typhoon in China this year, entered the mainland on July 28. It affected more than 720,000 people in Fujian province, the first coastal touch point, and caused a direct economic loss of CNY52.3 million (USD7.3 million), according to data released by the provincial flood control agency the same day.

Fengwu was released by the Shanghai AI Lab and other research institutes including the University of Science and Technology of China in April. The model, which is based on multimodal neural networks and multi-task learning methods, can simulate complicated weather activity.

Reducing error by one km in 24 hours can lower about CNY97 million in direct economic loss so accurate typhoon forecasting is vital in minimizing risks, a researcher said.

Editors: Dou Shicong, Emmi Laine