analiza aiqbench
This research introduces SoftSEEPS, a differentiable approximation of the SEEPS score that enables direct end-to-end training of machine learning models for precipitation forecasting. By optimizing on this metric alongside RMSE within a joint objective, the approach improves forecast accuracy without requiring post-processing calibration steps. This methodological advance represents a significant step forward in applying deep learning to geophysical prediction tasks.
SoftSEEPS improves ML-based precipitation forecasting
arXiv:2610.09752v1 Announce Type: new Abstract: In this paper we have developed a differentiable approximation of the well-known SEEPS score, which we name SoftSEEPS. This allows the training of a Machine Learning model to forecast precipitation directly. We test SoftSEEPS on the IMERG dataset (0.1 degree resolution) by training a decoder for precipitation on the latent space of a pre-trained low-resolution forecasting model. Combining SoftSEEPS and RMSE in a joint objective is possible with marginal trade-offs in either metric.
Abstract:In this paper we have developed a differentiable approximation of the well-known SEEPS score, which we name SoftSEEPS. This allows the training of a Machine Learning model to forecast precipitation directly. We test SoftSEEPS on the IMERG dataset (0.1 degree resolution) by training a decoder for precipitation on the latent space of a pre-trained low-resolution forecasting model. Combining SoftSEEPS and RMSE in a joint objective is possible with marginal trade-offs in either metric.
Machine Learning (cs.LG)
arXiv:2610.09752 [cs.LG]
(or arXiv:2610.09752v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.09752 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 167 cuvinte
Subiecte conexe
Introducing WeatherNext 3, our most advanced and accurate global weather AI model
Google DeepMind · 3 sept. 2026
Improving precipitation forecasts in an AI weather model using observational data
arXiv · 4 sept. 2026
Diffusion Precipitation Generation with Medium-range Weather Forecasting
arXiv · 12 oct. 2026
ORBIT-2 based Weather and Climate Downscaling and Downscaled Global Forecasts on AMD Instinct
AMD · 8 iun. 2026