DeepMind’s hurricane model outran the forecasters — and even its builders can’t explain why
Google DeepMind's WeatherNext AI model has demonstrated superior accuracy in predicting hurricane intensity and track compared to traditional forecasting methods, outperforming them by approximately one day. Researchers are surprised by its effectiveness, especially given its use of lower-resolution atmospheric data, and are working to understand the underlying mechanisms of this 'black box' system. The model's ability to generate a wider spread of potential storm scenarios offers valuable lead time for disaster preparedness.
- Entities
- Topic: Artificial Intelligence
- Company: Google DeepMind
- Person: Mike Brennan
- Topic: Weather Forecasting
- Topic: Hurricanes
- Company: Google Research
- Person: Kate Musgrave
- Person: Ferran Alet
This article was reported and published by egamers.io. feed.gg links to it as part of a story cluster — full text, images and rights remain with the publisher.