Latent Gaussian Processes for Distribution Estimation of Multivariate Categorical Data

Multivariate categorical data occur in many applications of machine learning. One of the main difficulties with these vectors of categorical variables is sparsity. The number of possible observations