Package gov.sandia.cognition.graph.inference

Interface Summary

Interface	Description
EnergyFunction<LabelType>	This interface defines the methods necessary for an object to be passed into an iterative message-passing data inferencing solver (such as belief propagation).
EnergyFunctionSolver<LabelType>	Interface for any belief propagation implementation.
GraphWrappingEnergyFunction.PotentialHandler<LabelType,NodeNameType>	This interface defines the problem-specific methods this class requires as input
NodeNameAwareEnergyFunction<LabelType,NodeNameType>	Helpful interface that adds per-node labeling and getting results from energy functions using the graph's node label type.

Class Summary

Class	Description
CostSpeedupEnergyFunction<LabelType,NodeNameType>	This class trades memory usage (to store all of the costs) for compute time.
EdgeMergingEnergyFunction<LabelType,NodeNameType>	Our implementation of belief propagation requires that there be at most one edge between any pair of nodes.
GraphWrappingEnergyFunction<LabelType,NodeNameType>	This class is a simple wrapper for any input graph becoming a BP-capable graph.
SumProductBeliefPropagation<LabelType>	This class implements the sum-product belief propagation algorithm for arbitrary energy functions.
SumProductDirectedPropagation<LabelType>	This class implements a Bayes Net -- but only allowing for pairwise influence along edges.
SumProductInferencingAlgorithm<LabelType>	Base class for Sum-Product inferencing algorithms on graphs/energy functions