Exploiting Evidence in Probabilistic Inference

TL;DR

This paper introduces a logical processing-based method for evidence compilation in Bayesian networks, improving inference efficiency in various applications like genetic analysis and noisy-or networks.

Contribution

It presents a novel evidence-based compilation approach that enhances probabilistic inference, applicable to non-deterministic networks and outperforming existing algorithms.

Findings

Effective in genetic linkage analysis

Outperforms quickscore in noisy-or networks

Applicable to non-deterministic Bayesian networks

Abstract

We define the notion of compiling a Bayesian network with evidence and provide a specific approach for evidence-based compilation, which makes use of logical processing. The approach is practical and advantageous in a number of application areas-including maximum likelihood estimation, sensitivity analysis, and MAP computations-and we provide specific empirical results in the domain of genetic linkage analysis. We also show that the approach is applicable for networks that do not contain determinism, and show that it empirically subsumes the performance of the quickscore algorithm when applied to noisy-or networks.