Google’ın Yapay Zeka Tabanlı Hava Durumu Modeli, Dünyanın En İyi Tahmin Sistemlerini Geride Bırakıyor

As one of the leading companies in artificial intelligence, Google aimed to apply this technology to weather forecasting. Recently, GenCast, developed by DeepMind, was announced in this context.

Trained with 40 years of weather data, GenCast analyzes the information and is capable of making accurate predictions for up to 15 days within a very short 8-minute timeframe.

Better than the world’s top weather forecasting system

In its statements regarding GenCast, Google indicated that comparisons were made with existing advanced systems. Accordingly, it has shown better performance than the European Centre for Medium-Range Weather Forecasts’ ENS system, which is regarded as the most advanced weather forecasting system in the world.

The tests conducted were based on weather predictions for the year 2019. The results revealed that GenCast, which was trained using weather data up to 2018, provided 97.2% more accurate results compared to ENS.

Of course, it is not yet a perfect model. It is still in the development phase and has some flaws. However, it has the potential to become a system that could genuinely save lives in the future. It’s noteworthy that GenCast is just one of Google’s weather prediction models; the company plans to integrate these models into its services like Search and Maps in the upcoming period.