A generalized Wasserstein-2 loss with a differentiable surrogate lets stochastic neural networks reconstruct random field models whose outputs combine continuous and categorical variables.
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
CONDITIONAL 1representative citing papers
citing papers explorer
-
A generalized Wasserstein-2 distance approach for efficient reconstruction of random field models using stochastic neural networks
A generalized Wasserstein-2 loss with a differentiable surrogate lets stochastic neural networks reconstruct random field models whose outputs combine continuous and categorical variables.