A meteorological prediction system based on DeepMind’s artificial intelligence has managed to make precise rain forecasts with a period of one or two hours in advance instead of a few days as current numerical prediction models do.
The numerical prediction models of time are able to offer predictions on a planetary scale with several days of advancement, but fail to make predictions in short periods of time, a few hours.
Knowing accurately if it is going to rain one or two times it could help you organize better day by day, and even has a direct impact on water management, agriculture, aviation, emergency planning and outdoor events
.
This obstacle is the one who seeks to improve the Deepmind team, in collaboration with the National Meteorological Service of the United Kingdom, to offer accurate predictions under deadlines under two hours.
The immediate forecasts are possible thanks to the radar data and in combination with automatic learning.
In adopted approach, it uses statistical, economic and cognitive measures in a generative model that performs predictions based on previous radars.
“Our model produces realistic and consistent predictions spacing on regions of up to 1,536 km x 1,280 kilometers and with delivery times between 5 and 90 minutes in advance,” they explain in the text published in Nature.
Researchers ensure that the generative model has first classified in the evaluations of more than 50 meteorologists due to their “accuracy and utility in 89 percent of cases against two competitive methods.”
