Approximate Inference via Weighted Rademacher Complexity.
Jonathan KuckAshish SabharwalStefano ErmonPublished in: AAAI (2018)
Keyphrases
- approximate inference
- rademacher complexity
- graphical models
- data dependent
- belief propagation
- probabilistic inference
- parameter estimation
- message passing
- bayesian networks
- gaussian process
- risk bounds
- latent variables
- dynamic bayesian networks
- conditional random fields
- least squares
- probabilistic model
- generalization error
- higher order
- maximum likelihood
- em algorithm
- decision trees