Copulas as High-Dimensional Generative Models: Vine Copula Autoencoders.
Natasa TagasovskaDamien AckererThibault VatterPublished in: NeurIPS (2019)
Keyphrases
- generative model
- high dimensional
- dependence structure
- sparse coding
- probabilistic model
- deep belief networks
- denoising
- mixture model
- discriminative models
- prior knowledge
- dimensionality reduction
- low dimensional
- maximum entropy principle
- em algorithm
- hierarchical hidden markov models
- conditional random fields
- data points
- discriminative learning
- marginal distributions
- hidden variables
- image processing
- object categories
- topic models
- graphical models
- semi supervised
- feature space
- markov networks
- deep learning
- generative and discriminative models
- parameter learning
- semi supervised learning
- expectation maximization
- representational power
- support vector
- information retrieval