GoogleThe company's latest release is the new NeuralGCM AI Modelsthat combines machine learning and traditional techniques to build new AI weather prediction model,The results were published yesterday inNatureJournals.
Google says NeuralGCM is characterized by lower cost and higher accuracy in forecasting weather 1-10 days into the future than other weather forecasting models based purely on machine learning.
The NeuralGCM model is open source and can be run relatively quickly by users on laptops, says study co-author Stephen Hoyer of Google Research.
The NeuralGCM model was developed by Google Research, Google DeepMindScientists at MIT, Harvard University and the European Center for Medium-Range Weather Forecasts (ECMWF) have developed it using machine learning and neural networks, and IT Home has attached its training model below:
-
Using brain neurons as a model, trained on weather data from the past few decades.
-
It also uses physical equations to describe large-scale weather patterns, essentially combining GCMs, physics-intensive methods with AI-driven tasks.
-
Some large-scale physics was retained and artificial intelligence was used to replace some of the modeling.
The result, the researchers say, is that the model can make high-quality forecasts faster with less computing power. They say NeuralGCM is as accurate as the European Center for Medium-Range Weather Forecasts (ECMWF) 1-15 day forecasts.
Attach reference address