SEEDS, Google's new AI to predict the weather reliably

SEEDS from Google AI to predict the weather

Google has implemented artificial intelligence in the field of weather forecasts with the aim of optimizing weather prediction and studying the consequences of climate change. This will be done with SEEDS, your project to generate accurate and economic forecasts, for a better understanding of the climate. Let's see how it works and what potential this new Google project has.

Google is working on the development of SEEDS

Improvement of the traditional meteorological system

SEEDS, name formed by the acronym of Scalable Ensemble Envelope Diffusion Sampler, is the latest technology based on artificial intelligence generative developed by Google and that aims to improve weather prediction.

This technology takes traditional weather forecast data to teach artificial intelligence how inclement weather develops and thus be able to give a more reliable and more economical forecast. The development of this technology We can see it in the scientific papers that Google has presented to the scientific journal ScienceAdvances.

The SEEDS project is therefore proposed as the most reliable current tool to assess the risks of climate change. As I say, it does so through the interpretation of these climate projections based on artificial intelligence, which are even capable of proposing different weather scenarios and calculate the probabilities of these variations occurring.

For this reason, a fundamental part of this research is the measurement of uncertainty in the climate. And today, the climate prediction models we have do not offer probabilities of the different meteorological scenarios that may occur. With this advance It is also intended to understand the possible real consequences of climate change., which leads us to wonder why this technology has not been developed before.

Why hasn't it been done before?

accurate weather forecasts

Currently, to study meteorology and climate changes we need to generate sets of physics-based forecasts. This is the traditional way of predicting weather and is very expensive in terms of data computing. Thus, Google's proposal is to reduce costs using models, based on generative artificial intelligence, that have learned the entire historical database on the climate of a specific area.

This data makes it possible to emulate inclement weather and its evolution while consuming only a fraction of the resources traditionally spent. This translates into the ability to generate realistic forecasts in a much more economical way. And it not only implies that the process is less expensive but also leads us to a greater reliability and precision in prediction of these meteorological events.

So basically it hasn't been done before because the cost of the computational calculation was too high. Now it can be done thanks to the artificial revolution we are experiencing. It undoubtedly represents an impressive advance in the field of meteorology.

SEEDS represents a breakthrough in weather prediction

Most reliable weather forecast

It is not surprising that SEEDS represents a turning point in climatological study and that this advance represents the beginning of many improvements in weather prediction. In fact, I am going to give you something so that you understand the magnitude of the calculations that this technology allows and that is that SEEDS is capable of assessing around 114 million parameters in one sitting.

Now, we still cannot enjoy this tool since This project is still in development and it has not yet been definitively released. Although at the moment we are not very worried since we have good applications to check the weather.

Ultimately, we will have to wait for Google to bring us its weather prediction tool. Knowing that the scientific papers arrived at the end of last month we can understand that until the end of the year or even next year, we will not have this tool.