Flexible and Efficient Inference with Particles for the Variational Gaussian Approximation.
Théo Galy-FajouValerio PerroneManfred OpperPublished in: Entropy (2021)
Keyphrases
- efficient inference
- probabilistic inference
- fully connected
- approximate inference
- markov random field
- conditional random fields
- hidden variables
- graph structure
- variational methods
- image segmentation
- markov networks
- exact inference
- closed form
- graph cuts
- higher order
- linear models
- probabilistic model
- structured prediction
- human pose estimation
- computer vision
- graphical models
- belief propagation
- message passing
- optical flow
- factor graphs
- prior knowledge