Variational Gaussian Copula Inference
This work addresses the problem of efficient inference in complex Bayesian models for researchers and practitioners, representing an incremental improvement by combining existing copula and variational methods.
The paper tackled the challenge of constructing flexible variational proposals for hierarchical Bayesian models with non-Gaussian hidden variables, proposing a semiparametric variational Gaussian copula approach that preserves multivariate dependence and achieves competitive performance on benchmark datasets.
We utilize copulas to constitute a unified framework for constructing and optimizing variational proposals in hierarchical Bayesian models. For models with continuous and non-Gaussian hidden variables, we propose a semiparametric and automated variational Gaussian copula approach, in which the parametric Gaussian copula family is able to preserve multivariate posterior dependence, and the nonparametric transformations based on Bernstein polynomials provide ample flexibility in characterizing the univariate marginal posteriors.