M. Pawan Kumar 
HOME

This course has been replaced by a new MVA course. SLIDES
Lecture 1: Probabilistic Models.
Lecture 2: Belief Propagation.
Lecture 3: Algorithms based on Minimum Cut.
Lecture 4, Part 1: Convex Relaxations.
Lecture 4, Part 2: LP Relaxation and its Dual.
Lecture 5: Tree Reweighted Message Passing, Dual Decomposition.
Lecture 6, Part 1: Dual Decomposition.
Lecture 6, Part 2: Belief Propagation for Computing Marginals.
Lecture 7: Free energy approximations. EXTERNAL LINKS
The following paper provides speedups for special cases of belief propagation.
More speedups for special cases of belief propagation.
The following paper provides the graph construction for an energy function defined using 2 labels for each random variable.
The following paper describes a minimum cut based interactive image segmentation system.
The following paper describes the expansion algorithm.
The following thesis contains the derivation of the multiplicative bound for the expansion algorithm (section 4.3.4).
The following paper provides a comparison of different convex relaxations.
The following paper provides the treebased dual of the LP relaxation and the original TRW algorithm.
The following paper describes the sequential TRW algorithm. It also provides a brief description of belief propagation as reparameterization (subsection 2.2).
The following paper describes the dual decomposition algorithm.
The following paper describes the free energy approximations.
The following paper provides the messages for generalized belief propagation. RESOURCES Energy minimization benchmark for smoothnessbased priors.
Tutorial on (sumproduct treereweighted) message passing algorithms.
Tutorial on LP relaxation.
Online course. 