You will no longer complain about the weather forecast, because Google has created an AI capable of predicting it with great precision better than current models.
One of the most difficult things to predict are the weather forecasts, and although today with new technologies they are quite close, at some point you will have looked up at the sky, and you will have seen that it was sunny when they had said it was going to rain. .
Luckily Google has created a technology capable of generate accurate weather forecasts at scale and is cheaper than current forecasts and methods.
Scalable Ensemble Envelope Diffusion Sampler (SEEDS) acts similar to large language models like ChatGPT.
Its technology generates many more sets or multiple weather scenarios faster and cheaper than traditional models.
On the one hand we have the physics-based predictions currently used by the weather services we know, which compile several measurements and give a final prediction that averages many different modeled predictions.
These current technologies also use deterministic or probabilistic forecasting models where random variables are introduced in the initial conditions, but this leads to a higher error rate, making it difficult to accurately predict extreme weather.
The Advantages of Scalable Ensemble Envelope Diffusion Sampler (SEEDS)
However Scalable Ensemble Envelope Diffusion Sampler (SEEDS) by Google produces forecast models from physical measurements collected by meteorological agencies, analyzing the relationships between the unit potential energy per mass of the Earth’s gravitational field in the mid-troposphere and sea level pressure.
This artificial intelligence can extrapolate up to 31 sets of forecasts, based on only one or two forecasts used as input data.
On the other hand, they clarify that the computing costs associated with carrying out calculations with Scalable Ensemble Envelope Diffusion Sampler (SEEDS) are insignificant compared to current methods.
