Lifting Factor Graphs with Some Unknown Factors

TL;DR

This paper introduces the LIFAGU algorithm, which leverages symmetries in factor graphs with unknown factors to enable efficient and exact lifted probabilistic inference by transferring known potentials to unknown ones.

Contribution

The paper presents a novel algorithm that extends lifting techniques to factor graphs with unknown factors, facilitating exact inference in more complex models.

Findings

LIFAGU successfully identifies symmetric subgraphs with unknown factors.

The method enables transfer of known potentials to unknown factors for inference.

LIFAGU maintains exactness while improving inference efficiency.

Abstract

Lifting exploits symmetries in probabilistic graphical models by using a representative for indistinguishable objects, allowing to carry out query answering more efficiently while maintaining exact answers. In this paper, we investigate how lifting enables us to perform probabilistic inference for factor graphs containing factors whose potentials are unknown. We introduce the Lifting Factor Graphs with Some Unknown Factors (LIFAGU) algorithm to identify symmetric subgraphs in a factor graph containing unknown factors, thereby enabling the transfer of known potentials to unknown potentials to ensure a well-defined semantics and allow for (lifted) probabilistic inference.