SoftSEEPS improves ML-based precipitation forecasting

Jost Arndt, Utku Isil, Noelia Otero, Rodrigo Almeida, Wojciech Samek +1 more
10/7/2026
cs.LG

No confident code implementation found for this paper.

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.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

GenONet: A Generative operator Network for High-Resolution Precipitation Nowcasting

86% match2026

Knowledge-Guided Adaptive Mixture of Experts for Precipitation Prediction

85% match2025

Self-Supervised Pre-Training for Precipitation Post-Processor

85% match2023

PostRainBench: A comprehensive benchmark and a new model for precipitation forecasting

84% match1 repo292023

Deep learning for precipitation nowcasting: A survey from the perspective of time series forecasting

84% match2024

Probabilistic Weather Forecasting with Hierarchical Graph Neural Networks

84% match2026

Cite this paper

@article{arndt2026softseeps,
  title  = {SoftSEEPS improves ML-based precipitation forecasting},
  author = {Jost Arndt and Utku Isil and Noelia Otero and Rodrigo Almeida and Wojciech Samek and Jackie Ma},
  year   = {2026},
  eprint = {2610.09752},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2610.09752v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2610.09752.svg)](https://codeofpaper.com/paper/2610.09752)

Discussion