### Resumé

to be Conditional Linear Gaussian (CLG). The use of AR eliminates the need for complex matrix operations in case of CLG networks, while offering opportunities to exploit additional independence and irrelevance properties in both cases when compared to Variable

Elimination (VE). The performance of Simple Propagation with AR has been evaluated on a set of real-world Bayesian networks with discrete variables and hybrid Bayesian networks constructed by randomly replacing discrete variables with continuous variables under the CLG constraints. The performance of Simple Propagation with AR is compared with the performance of Lazy Propagation with AR. The results of the experimental performance analysis of Simple Propagation with AR are encouraring

Originalsprog | Dansk |
---|---|

Titel | Proceedings of Machine Learning Research |

Vol/bind | 72 |

Forlag | PMLR Press |

Publikationsdato | 2018 |

Sider | 260-271 |

Status | Udgivet - 2018 |

Begivenhed | International Conference on Probabilistic Graphical Models - Varighed: 11 sep. 2018 → 14 sep. 2018 http://pgm2018.utia.cz |

### Konference

Konference | International Conference on Probabilistic Graphical Models |
---|---|

Periode | 11/09/2018 → 14/09/2018 |

Internetadresse |

Navn | The Proceedings of Machine Learning Research |
---|---|

ISSN | 2640-3498 |

### Citer dette

*Proceedings of Machine Learning Research*(Bind 72, s. 260-271). PMLR Press. The Proceedings of Machine Learning Research

}

*Proceedings of Machine Learning Research.*bind 72, PMLR Press, The Proceedings of Machine Learning Research, s. 260-271, International Conference on Probabilistic Graphical Models, 11/09/2018.

**Simple Propagation with Arc-Reversal in Bayesian Networks.** / Madsen, Anders Læsø; Butz, Cory J.; Oliveira, Jhonatan; dos Santos, Andre E.

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Konferenceartikel i proceeding › Forskning › peer review

TY - GEN

T1 - Simple Propagation with Arc-Reversal in Bayesian Networks

AU - Madsen, Anders Læsø

AU - Butz, Cory J.

AU - Oliveira, Jhonatan

AU - dos Santos, Andre E.

PY - 2018

Y1 - 2018

N2 - Simple Propagation is a recently introduced algorithm for inference in discrete Bayesian networks using message passing in a junction tree. Simple Propagation is similar to Lazy Propagation, but uses the simple one in, one out-principle when computing messages between cliques of the junction tree instead of using a more in-depth graphical analysis of the set of potentials. In this paper, we describe how to apply Arc-Reversal (AR) as the marginalization algorithm during message passing in Simple Propagation. We consider both discrete and hybrid Bayesian networks, where the continuous variables are assumedto be Conditional Linear Gaussian (CLG). The use of AR eliminates the need for complex matrix operations in case of CLG networks, while offering opportunities to exploit additional independence and irrelevance properties in both cases when compared to VariableElimination (VE). The performance of Simple Propagation with AR has been evaluated on a set of real-world Bayesian networks with discrete variables and hybrid Bayesian networks constructed by randomly replacing discrete variables with continuous variables under the CLG constraints. The performance of Simple Propagation with AR is compared with the performance of Lazy Propagation with AR. The results of the experimental performance analysis of Simple Propagation with AR are encouraring

AB - Simple Propagation is a recently introduced algorithm for inference in discrete Bayesian networks using message passing in a junction tree. Simple Propagation is similar to Lazy Propagation, but uses the simple one in, one out-principle when computing messages between cliques of the junction tree instead of using a more in-depth graphical analysis of the set of potentials. In this paper, we describe how to apply Arc-Reversal (AR) as the marginalization algorithm during message passing in Simple Propagation. We consider both discrete and hybrid Bayesian networks, where the continuous variables are assumedto be Conditional Linear Gaussian (CLG). The use of AR eliminates the need for complex matrix operations in case of CLG networks, while offering opportunities to exploit additional independence and irrelevance properties in both cases when compared to VariableElimination (VE). The performance of Simple Propagation with AR has been evaluated on a set of real-world Bayesian networks with discrete variables and hybrid Bayesian networks constructed by randomly replacing discrete variables with continuous variables under the CLG constraints. The performance of Simple Propagation with AR is compared with the performance of Lazy Propagation with AR. The results of the experimental performance analysis of Simple Propagation with AR are encouraring

M3 - Konferenceartikel i proceeding

VL - 72

T3 - The Proceedings of Machine Learning Research

SP - 260

EP - 271

BT - Proceedings of Machine Learning Research

PB - PMLR Press

ER -