Google says its new artificial intelligence model outperforms top weather forecasting systems

Google's DeepMind team launched a weather forecast artificial intelligence model called GenCast this week.

exist Papers published in NatureDeepMind researchers said they found that GenCast outperformed the European Center for Medium-Range Weather Forecasts' ENS – apparently the world's top commercial forecasting system.

and in a blog postthe DeepMind team provided a more understandable explanation of the technology: While its previous weather models were “deterministic and provided a single, best estimate of future weather,” GenCast “contains 50 or more predictions “A collection of forecasts, each representing a possible weather trajectory,” creating a “complex probability distribution of future weather scenarios.”

As for how it compares to ENS, the team said they trained GenCast using weather data up to 2018, then compared the forecasts for 2019 and found that GenCast was more accurate 97.2% of the time.

Google said GenCast is part of its suite of artificial intelligence-based weather models that are starting to be integrated into Google Search and Maps. It also plans to publish GenCast's immediate and historical forecasts, which anyone can use in their own research and models.