A multidimensional Gaussian loss with Fourier-represented covariances and an information-sharing regularizer lets a network emit closed-form correlated predictive distributions, demonstrated on wind-speed super-resolution.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multidimensional Distributional Neural Network Output Demonstrated in Super-Resolution of Surface Wind Speed
A multidimensional Gaussian loss with Fourier-represented covariances and an information-sharing regularizer lets a network emit closed-form correlated predictive distributions, demonstrated on wind-speed super-resolution.