NASA And IBM Develop AI Basic Model For Climate And Weather Applications

NASA and IBM created a basic AI model with geospatial satellite data (photo: dock. NASA)

JAKARTA - The United States Aeronautics and Space Agency (NASA) is working with IBM to build an application that combines Artificial Science and Intelligence (AI).

From the Engadget report, the applications to be developed by NASA and IBMakan are used to measure weather and climate. All Earth sciences will be added to the application so that the technology can work optimally.

GraphCast and Fourcastnet are examples of AI models used to forecast weather. Although its performance is faster than traditional models, IBM says that both are AI emulators.

The software is certainly different from the basic AI models that NASA and IBM want to develop. The AI emulator can only predict the weather based on training datasets, while the basic models supporting the AI are generatively capable of encoding physics as the core weather forecast.

For now, neither IBM nor NASA, have detailed details of the apps to be created. They only share several goals such as wider accessibility, faster inference times, and greater diversity.

Another goal of constructing basic models for applications is more accurate climate forecasts in other areas such as meteorological phenomena predictions, identifying conducive conditions for aircraft turbulence, to forest fires.

To achieve this goal, this AI fundamental model will utilize data from NASA satellites. The data used is related to geospatials, namely data in the form of geographical locations, object characteristics, and other positions related to the earth.