Cutting Forecast Time: Aardvark Weather AI Predicts Weather in a Flash

Built through a joint effort by researchers from the UK and Canada, the system avoids traditional weather models that rely on heavy computing and mathematical simulations. Instead, it uses a type of machine learning to quickly create weather forecasts.

The project involved teams from the University of Cambridge, the Vector Institute at the University of Toronto, and the Alan Turing Institute. Their findings were published in the journal Nature.

Did you know?

Want to get smarter & wealthier with crypto?

Subscribe - We publish new crypto explainer videos every week!

What Does Staking Mean in Crypto? (Easily Explained!)

SUBSCRIBE

ON YOUTUBE

Unlike older systems that use equations to simulate the atmosphere, Aardvark Weather works by learning from large sets of data. It can generate detailed forecasts for wind, humidity, temperature, and air pressure on a global scale.

It also provides local-level reports, such as ground-level temperatures and wind speeds. The system handles this using three main parts: an encoder to read the data, a processor to analyze it, and a decoder to turn it into forecast information.

Using four NVIDIA A100 GPUs, Aardvark can complete a full global forecast in about one second. This faster process allows for more frequent updates and easier fine-tuning.

Aardvark Weather could also help places that cannot afford expensive weather systems because it requires less computing power.

Recently, Flok Health introduced an AI-powered app to help people in the UK manage back pain. How does it work? Read the full story.