-
Generated Dataset
The dataset is generated from the following model: X1:10, Y1:10 ∼ Uniform(0, 1), S ∼ Multinomial(2, Normal((1/2)X1, (1/2)X2)), Z | X, Y, S ∼ Uniform(0, X) Uniform(X, Z) -
MNIST, SVHN, and CelebA
The MNIST, SVHN, and CelebA datasets are used for conditional density estimation tasks. -
UCI Regression Datasets
The dataset used in the paper is the UCI Regression Datasets, which are a standard benchmark for evaluating conditional density estimators.