MMAF-guided learning constrains neural network training with Ornstein-Uhlenbeck process structure to generate calibrated spatio-temporal ensemble forecasts, where shallow feed-forward networks perform comparably to or better than convolutional or diffusion models.
Ncep cpc pentad olr anomalies on a 0.25° lat/lon grid, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Spatio-temporal probabilistic forecast using MMAF-guided learning
MMAF-guided learning constrains neural network training with Ornstein-Uhlenbeck process structure to generate calibrated spatio-temporal ensemble forecasts, where shallow feed-forward networks perform comparably to or better than convolutional or diffusion models.