Closed-form contrast-function approximations to log and exp maps on Fisher-Rao statistical manifolds yield fast wrapped Gaussian posteriors that capture skewness and heavy tails.
International Conference on Artificial Intelligence and Statistics (AISTATS) , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Beyond Laplace: Closed-form wrapped Gaussian posterior approximations on statistical manifolds
Closed-form contrast-function approximations to log and exp maps on Fisher-Rao statistical manifolds yield fast wrapped Gaussian posteriors that capture skewness and heavy tails.