BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling.
Lars MaaløeMarco FraccaroValentin LiévinOle WintherPublished in: CoRR (2019)
Keyphrases
- latent variables
- probabilistic model
- posterior distribution
- latent variable models
- random variables
- prior knowledge
- gaussian process
- discriminative methods
- hierarchical model
- topic models
- generative model
- real valued
- approximate inference
- bayesian networks
- probabilistic graphical models
- max margin
- variational bayes
- observed variables
- conditionally independent
- structured prediction
- gaussian processes
- graphical models
- structured output
- pairwise