Localized Partial Evaluation of Belief Networks

TL;DR

The paper introduces the Localized Partial Evaluation (LPE) algorithm that efficiently computes probability bounds in belief networks by focusing only on relevant network parts, offering an anytime approximation method.

Contribution

It presents a novel localized partial evaluation algorithm for belief networks that provides interval bounds and improves with additional computation time.

Findings

LPE computes interval bounds efficiently for large networks.

LPE's accuracy improves with more computation time.

LPE is suitable for applications needing approximate probabilities.

Abstract

Most algorithms for propagating evidence through belief networks have been exact and exhaustive: they produce an exact (point-valued) marginal probability for every node in the network. Often, however, an application will not need information about every n ode in the network nor will it need exact probabilities. We present the localized partial evaluation (LPE) propagation algorithm, which computes interval bounds on the marginal probability of a specified query node by examining a subset of the nodes in the entire network. Conceptually, LPE ignores parts of the network that are "too far away" from the queried node to have much impact on its value. LPE has the "anytime" property of being able to produce better solutions (tighter intervals) given more time to consider more of the network.