The project contains an implementation of Loopy Belief Propagation, a popular message passing algorithm for performing inference in probabilistic graphical models. It provides exact inference for graphical models without loops. While inference for graphical models with loops is approximate, in practice it is shown to work well. Our implementation is generic and operates on factor graph representation of graphical models. It handles factors of any order, and variable domains of any size. It is implemented in Apache Spark GraphX, and thus can scale to large scale models. Further, it supports computations in log scale for numerical stability. ###Examples
API
import sparkle.graph._
// Load graph from the inputPath
val graph = Utils.loadLibDAIToFactorGraph(sc, inputPath)
// Run Belief Propagation on the graph with maximum iterations "maxIterations" and epsilon "epsilon"
val beliefs = BeliefPropagation(graph, maxIterations, epsilon)
Spark submit
./spark-submit --master spark://MASTER:PORT --jars belief-propagation.jar --class sparkle.graph.BP INPUT_PATH OUTPUT_PATH N_ITERATIONS EPSILON
###References
- Belief propagation algorithm
- Yedidia, J.S.; Freeman, W.T.; Weiss, Y., "Understanding Belief Propagation and Its Generalizations" in Exploring Artificial Intelligence in the New Millennium, Lakemeyer, G. and Nebel, B., Eds., ISBN: 1-55860-811-7, chapter 8, pp. 239-236, Morgan Kaufmann Publishers, January 2003. (pdf) [An excellent description of BP]
- LibDAI file format is used with explicit factor ids: example.
- Talk by the project authors: Alexander Ulanov, Manish Marwah "Malicious site detection with large scale belief propagation", Strata+Hadoop, March 2017
###How to contribute Contributors are required to sign the Corporate Contributor License Agreement