Quasi Black-Box Variational Inference with Natural Gradients for Bayesian Learning.
Martin MagrisMostafa ShabaniAlexandros IosifidisPublished in: CoRR (2022)
Keyphrases
- black box
- bayesian learning
- posterior distribution
- variational inference
- probability distribution
- latent variables
- parameter estimation
- hyperparameters
- bayesian framework
- posterior probability
- bayesian inference
- maximum a posteriori
- markov chain monte carlo
- topic models
- model selection
- closed form
- test cases
- mixture model
- maximum likelihood
- gaussian distribution
- expectation maximization
- bayesian networks
- variational methods
- probabilistic model
- approximate inference
- gaussian process
- random sampling
- latent dirichlet allocation
- em algorithm
- cross validation
- least squares
- state space
- data mining