Importance Sampling via Variational Optimization

TL;DR

This paper introduces a variational optimization-based importance sampling algorithm for large Bayesian networks, improving convergence and accuracy in challenging scenarios with deterministic tables and unlikely observations.

Contribution

It presents a novel importance sampling method that adaptively updates the proposal distribution using variational techniques to enhance convergence in complex Bayesian networks.

Findings

Demonstrated improved convergence on genetic linkage analysis networks

Effectively handles deterministic conditional probability tables

Outperforms traditional sampling methods in difficult scenarios

Abstract

Computing the exact likelihood of data in large Bayesian networks consisting of thousands of vertices is often a difficult task. When these models contain many deterministic conditional probability tables and when the observed values are extremely unlikely even alternative algorithms such as variational methods and stochastic sampling often perform poorly. We present a new importance sampling algorithm for Bayesian networks which is based on variational techniques. We use the updates of the importance function to predict whether the stochastic sampling converged above or below the true likelihood, and change the proposal distribution accordingly. The validity of the method and its contribution to convergence is demonstrated on hard networks of large genetic linkage analysis tasks.