MOPED: Efficient priors for scalable variational inference in Bayesian deep neural networks.
Ranganath KrishnanMahesh SubedarOmesh TickooPublished in: CoRR (2019)
Keyphrases
- variational inference
- neural network
- bayesian inference
- bayesian models
- posterior distribution
- topic models
- gaussian process
- prior information
- bayesian framework
- bayesian networks
- mixture model
- maximum likelihood
- latent dirichlet allocation
- soft computing
- hyperparameters
- em algorithm
- dirichlet process
- probabilistic model