libDAI A free and open source C++ library for discrete approximate inference in graphical models
Download

libDAI Ranking & Summary

libDAI Tags

library approximate interference graphical models approximate inference factor graph

libDAI Description

A free and open source C++ library for discrete approximate inference in graphical models libDAI is a free and open source C++ development library that provides implementations of various (approximate) inference methods for discrete graphical models. libDAI supports arbitrary factor graphs with discrete variables; this includes discrete Markov Random Fields and Bayesian Networks.libDAI is targeted at researchers. To be able to use libDAI, a good understanding of graphical models is needed.The best way to use libDAI is by writing C++ code that invokes the library; in addition, part of the functionality is accessibly by using the:· command line interface· MatLab interface · python interface· octave interface.libDAI can be used to implement novel (approximate) inference algorithms and to easily compare the accuracy and performance with existing algorithms that have been implemented already.libDAI is a cross-platform library which works under Mac OS X, Linux and Windows. Here are some key features of "libDAI": · Exact inference by brute force enumeration; · Exact inference by junction-tree methods; · Mean Field; · Loopy Belief Propagation · Fractional Belief Propagation · Tree-Reweighted Belief Propagation · Tree Expectation Propagation · Generalized Belief Propagation · Double-loop GBP · Various variants of Loop Corrected Belief Propagation · Gibbs sampler · Conditioned Belief Propagation What's New in This Release: New features: · Implemented Tree-Reweighted BP · Implemented Fractional Belief Propagation Miscellaneous improvements: · Improved documentation · Added constructors for BBPCostFunction · Added accessor/mutator for the elements of a SmallSet · Added unary minus to TProb · Added Hellinger distance to TProb::DistType · Two small patches for src/jtree.cpp and utils/createfg.cpp · tests/testdai option "marginals" now has three possible values: NONE · outputs no marginals, VAR only variable marginals, and ALL outputs all · marginals calculated by the algorithm. · Strengthened convergence criteria of various algorithms · Added GraphAL, an adjacency list implementation for graphs, · similar to (but simpler than) BipartiteGraph · Added some functionality to create various standard factors · (the functions whose names start with "createFactor") · Added examples example_sprinkler_gibbs and example_sprinkler_em · Implemented various heuristics for choosing a variable elimination sequence · in JTree · Added BETHE method for GBP/HAK cluster choice · Made alias support official part of the library Build system: · Moved platform independent build options into Makefile.ALL · Windows platform now needs Boost 1.37 or higher Code cleanup: · Cleaned up utils/createfg and tests/testdai · Cleanup of some duplicate code in BP, FBP, TRWBP Interface changes (the old interface is marked obsolete and will be removed in the next release): · Removed obsolete/deprecated stuff · Renamed Graph into GraphEL (for Graph, implemented as Edge List) Renamed some functions of BipartiteGraph: · add1() -> addNode1() · erase1() -> eraseNode1() · nr1() -> nrNodes1() · add2() -> addNode2() · erase2() -> eraseNode2() · nr2() -> nrNodes2() Renamed some functions in ClusterGraph: · ClusterGraph::eliminationCost() -> ::eliminationCost_MinFill() · ClusterGraph::VarElim() -> ClusterGraph::VarElim( sequentialVariableElimination( ElimSeq ) ) · ClusterGraph::VarElim_MinFill() -> ClusterGraph::VarElim( greedyVariableElimination( eliminationCost_MinFill ) ) · Changed interface of one RegionGraph constructor · Renamed "Clamped BP" into "Conditioned BP" Bug fixes: · Fixed regression in scripts/regenerate-properties (now all errors are · gathered before throwing the exception, as used to be the case) · Fixed FTBFS errors on MACOSX (thanks to Sebastian Riedel for reporting this bug) · Fixed error in BP that occured with damping · and using log-domain updates · Fixed bug in HAK for trivial region graphs (with only one outer region · and no inner regions), reported by Alejandro Lage. · Fixed long-standing bug in TreeEP (now, within-loop propagation optimization works)

libDAI Related Software