We are concerned with algorithms for solving reasoning problems over graphical models, which includes common tasks for belief, constraint, and mixed networks. In the following we present some of the resulting work; this list is bound to grow over time.
See our Repository for some example problem instances.
Exact P(e)
aolibPE computes the exact probability of evidence for Bayesian networks using AND/OR search spaces.
Bucket_Elimination performs variable elimination to obtain the exact probability of evidence.
Approximate P(e)
IJGPIS computes approximate probability of evidence in Bayesian networks using importance sampling, Iterative Join Graph Propagation (IJGP) and Relational Consistency.
IJGPSampleSearch computes approximate probability of evidence in Bayesian networks using importance sampling, SampleSearch and Iterative Join Graph Propagation (IJGP).
Approximate Belief Updating
IJGP (Iterative Join Graph Propagation) computes approximate beliefs in Bayesian networks.
IJGPIS computes computes approximate beliefs in Bayesian networks using importance sampling, Iterative Join Graph Propagation (IJGP) and Relational Consistency.
IJGPSampleSearch computes approximate beliefs in Bayesian networks using importance sampling, SampleSearch and Iterative Join Graph Propagation (IJGP).
MPE
aolibMPE implements several exact AND/OR search algorithms to compute the most probable explanation in Bayesian networks using Mini-Bucket heuristics and caching.
mbeMPE implements the Mini-Bucket approximation of the most probable explanation in Bayesian networks.
WCSP
aolibWCSP finds the optimal solution to weighted constraint satisfaction problems via various AND/OR search algorithms using Mini-Bucket heuristics and caching.
mbeWCSP implements the Mini-Bucket approximation for weighted constraint satisfaction problems.
ILP
aolibILP computes the optimal solution to 0/1 integer linear programs using AND/OR search.
Tree Decomposition
treeDecomp computes a tree decomposition of a given graphical model.
Old software
REES offers a GUI to create and edit graphical models and implements a number of algorithms. It is no longer actively developed.
Disclaimer: All software should be regarded as in development and is made available on an "as-is" basis, without warranty of any kind. It has in no way been thoroughly tested and might not function as intended or expected.