Parameterized Complexity Results for Bayesian Inference

Publication date

2022-06-14

Authors

Bodlaender, HansORCID 0000-0002-9297-3330ISNI 0000000081342475
Donselaar, NilsISNI 0000000507286942
Kwisthout, Johan

Editors

Advisors

Supervisors

DOI

Document Type

/dk/atira/pure/researchoutput/researchoutputtypes/workingpaper/preprint
Open Access logo

License

cc_by

Abstract

We present completeness results for inference in Bayesian networks with respect to two different parameterizations, namely the number of variables and the topological vertex separation number. For this we introduce the parameterized complexity classes $\mathsf{W[1]PP}$ and $\mathsf{XLPP}$, which relate to $\mathsf{W[1]}$ and $\mathsf{XNLP}$ respectively as $\mathsf{PP}$ does to $\mathsf{NP}$. The second parameter is intended as a natural translation of the notion of pathwidth to the case of directed acyclic graphs, and as such it is a stronger parameter than the more commonly considered treewidth. Based on a recent conjecture, the completeness results for this parameter suggest that deterministic algorithms for inference require exponential space in terms of pathwidth and by extension treewidth. These results are intended to contribute towards a more precise understanding of the parameterized complexity of Bayesian inference and thus of its required computational resources in terms of both time and space.

Keywords

cs.CC

Citation

Bodlaender, H, Donselaar, N & Kwisthout, J 2022 'Parameterized Complexity Results for Bayesian Inference' arXiv, pp. 1-12. < https://doi.org/10.48550/arXiv.2206.07172 >