An Importance Sampling Algorithm Based on Evidence Pre-propagation

TL;DR

The paper introduces EPIS-BN, an importance sampling algorithm that improves Bayesian network inference under unlikely evidence by using evidence pre-propagation techniques, outperforming previous methods without requiring costly learning stages.

Contribution

It presents a novel importance sampling algorithm that combines loopy belief propagation and e-cutoff for efficient approximate importance function computation.

Findings

EPIS-BN outperforms AIS-BN on large real Bayesian networks.

EPIS-BN avoids the costly learning stage of AIS-BN.

EPIS-BN provides significant accuracy improvements in inference.

Abstract

Precision achieved by stochastic sampling algorithms for Bayesian networks typically deteriorates in face of extremely unlikely evidence. To address this problem, we propose the Evidence Pre-propagation Importance Sampling algorithm (EPIS-BN), an importance sampling algorithm that computes an approximate importance function by the heuristic methods: loopy belief Propagation and e-cutoff. We tested the performance of e-cutoff on three large real Bayesian networks: ANDES, CPCS, and PATHFINDER. We observed that on each of these networks the EPIS-BN algorithm gives us a considerable improvement over the current state of the art algorithm, the AIS-BN algorithm. In addition, it avoids the costly learning stage of the AIS-BN algorithm.