Google DeepMind has developed an artificial intelligence model that accurately predicts the weather for the next ten days

DeepMind

Google DeepMind has created a new GraphCast artificial intelligence model that can generate ten-day weather forecasts in less than a minute. The developers claim that it is as accurate as traditional numerical forecasting methods.

Here's What We Know

GraphCast is a graph neural network trained on 39 years of weather data. It analyses the interrelationships of various weather variables and builds a forecast on this basis.

Unlike traditional approaches, the DeepMind model uses data from satellites, radars and weather stations. At the same time, its training requires significant computational resources.

In testing, GraphCast outperforms the ECMWF reference system in forecast accuracy for most variables and forecast horizons. The model performed particularly well in analysing the troposphere.

However, the DeepMind model has shortcomings. Its forecasts can be incomplete and are also difficult at high altitudes. According to the developers, AI cannot yet fully replace traditional numerical forecasting methods.

Google DeepMind has published the GraphCast code on GitHub.

Source: Science