Imagine being able to foresee extreme weather events with the nearest accuracy more than 2 weeks prior. Well, Google’s DeepMind has changed this vision into a reality by introducing “GenCast”. In this article, we will talk in-depth about this new functionality.
GenCast overview
If you know there is a 75% chance of extreme rainfall in your area enough to prepare for potential damages. That’s the power of probabilistic forecasting. By combining machine learning with probabilistic forecasting GenCast — an AI model from Google can predict the weather 15 days in advance. Moreover, it does not only predict what might happen but also how likely it is to occur.
As per the report, GenCast even beats the European Centre for Medium-Range Weather Forecasts (an international organization that is considered to make the most accurate predictions for its member states). For such accuracy, this model has been trained on forty years of historical data from ECMWF’s archive which includes variables such as temperature, wind, speed, and pressure at various altitudes.
- Also, read:
What makes it different?
While the traditional weather model gave one “best guess” of what the weather would look like in the future, GenCast provides 50 or more different predictions. Some of these predictions might represent extreme weather events, such as storms or heatwaves, that are less likely but still possible. By incorporating these possibilities, we can better prepare for worst-case scenarios, even if they have very little chance of occurring.
GenCast’s Challenges
Google has trained its AI model on older datasets. So there are chances of performance errors that might occur in the long run due to climate change. Let’s hope this innovation does not fail.
- Meanwhile, watch our Vivo X200 Pro Unboxing video:
