Provable Gradient Variance Guarantees for Black-Box Variational Inference.
Justin DomkePublished in: NeurIPS (2019)
Keyphrases
- black box
- variational inference
- bayesian inference
- topic models
- posterior distribution
- probabilistic graphical models
- variational methods
- black boxes
- probabilistic model
- gaussian process
- latent dirichlet allocation
- mixture model
- closed form
- exponential family
- white box
- exact inference
- latent variables
- edge detection
- test cases
- multiscale
- approximate inference
- graphical models
- factor graphs
- prior information
- probability distribution
- data sets
- posterior probability
- hidden variables
- generative model