A Score-and-Search Approach to Learning Bayesian Networks with Noisy-OR Relations

Charupriya Sharma, Zhenyu Liao, James Cussens, Peter Beek

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

4 Citations (Scopus)

Abstract

A Bayesian network is a probabilistic graphical model that consists of a directed acyclic graph (DAG), where each node is a random variable and attached to each node is a conditional probability distribution (CPD). A Bayesian network can be learned from data using the well-known score-and-search approach, and within this approach a key consideration is how to simultaneously learn the global structure in the form of the underlying DAG and the local structure in the CPDs. Several useful forms of local structure have been identified in the literature but thus far the score-and-search approach has only been extended to handle local structure in form of context-specific independence. In this paper, we show how to extend the score-and-search approach to the important and widely useful case of noisy-OR relations. We provide an effective gradient descent algorithm to score a candidate noisy-OR using the widely used BIC score and we provide pruning rules that allow the search to successfully scale to medium sized networks. Our empirical results provide evidence for the success of our approach to learning Bayesian networks that incorporate noisy-OR relations.
Original languageEnglish
Title of host publicationProceedings of Machine Learning Research
Subtitle of host publicationProceedings of the 10th International Conference on Probabilistic Graphical Models
EditorsManfred Jaeger, Thomas D. Nielsen
Pages413-424
Number of pages12
Volume138
Publication statusPublished - 25 Sept 2020
Event10th International Conference on Probabilistic Graphical Models - Hotel Comwell Rebild Bakker, Skørping, Denmark
Duration: 23 Sept 202025 Sept 2020
https://pgm2020.cs.aau.dk/

Publication series

NameProceedings of Machine Learning Research
Volume138
ISSN (Electronic)2640-3498

Conference

Conference10th International Conference on Probabilistic Graphical Models
Abbreviated titlePGM 2020
Country/TerritoryDenmark
CitySkørping
Period23/09/2025/09/20
Internet address

Fingerprint

Dive into the research topics of 'A Score-and-Search Approach to Learning Bayesian Networks with Noisy-OR Relations'. Together they form a unique fingerprint.

Cite this