A Scheme for Approximating Probabilistic Inference

TL;DR

This paper introduces a flexible class of probabilistic approximation algorithms using bucket elimination, capable of balancing accuracy and efficiency for various inference tasks, with initial empirical validation.

Contribution

It presents a novel probabilistic approximation framework based on bucket elimination that allows adjustable accuracy and efficiency levels for inference tasks.

Findings

Algorithms offer adjustable accuracy and efficiency.

Preliminary empirical evaluation on random networks.

Identifies regions where the approximation is complete.

Abstract

This paper describes a class of probabilistic approximation algorithms based on bucket elimination which offer adjustable levels of accuracy and efficiency. We analyze the approximation for several tasks: finding the most probable explanation, belief updating and finding the maximum a posteriori hypothesis. We identify regions of completeness and provide preliminary empirical evaluation on randomly generated networks.