SoftSEEPS improves ML-based precipitation forecasting
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.
ProofPaper ↗
Key points
- 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.
- Combining SoftSEEPS and RMSE in a joint objective is possible with marginal trade-offs in either metric.
Sources (1)
- [1]SoftSEEPS improves ML-based precipitation forecastingarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 09:40 AM
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.
In this paper we have developed a differentiable approximation of the well-known SEEPS score, which we name SoftSEEPS.
Extractive summary: sentences quoted from the sources.