Composing graphical models with neural networks for structured representations and fast inference.
Matthew J. JohnsonDavid DuvenaudAlexander B. WiltschkoRyan P. AdamsSandeep R. DattaPublished in: NIPS (2016)
Keyphrases
- graphical models
- structured representations
- bayesian networks
- neural network
- probabilistic inference
- exact inference
- map inference
- belief propagation
- probabilistic model
- random variables
- approximate inference
- belief networks
- factor graphs
- efficient inference algorithms
- loopy belief propagation
- undirected graphical models
- probabilistic graphical models
- markov networks
- information extraction
- plan recognition
- directed graphical models
- structure learning
- structured documents
- conditional random fields
- knowledge base
- influence diagrams
- bayesian inference
- conditional probabilities
- possibilistic networks
- decision making