Learning Bayesian networks with low inference complexity

One of the main research topics in machine learning nowadays is the improvement of the inference and learning processes in probabilistic graphical models. Traditionally, inference and learning have been treated separately, but given that the structure of the model conditions the inference complexity...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Progress in artificial intelligence Ročník 5; číslo 1; s. 15 - 26
Hlavní autoři: Benjumeda, Marco, Larrañaga, Pedro, Bielza, Concha
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.02.2016
Témata:
Artificial Intelligence
Computational Intelligence
Computer Imaging
Computer Science
Control
Data Mining and Knowledge Discovery
Mechatronics
Natural Language Processing (NLP)
Pattern Recognition and Graphics
Regular Paper
Robotics
Vision
Structure learning
Bayesian networks
Probabilistic graphical models
Network polynomials
Arithmetic circuits
Thin models
ISSN:2192-6352, 2192-6360
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:One of the main research topics in machine learning nowadays is the improvement of the inference and learning processes in probabilistic graphical models. Traditionally, inference and learning have been treated separately, but given that the structure of the model conditions the inference complexity, most learning methods will sometimes produce inefficient inference models. In this paper we propose a framework for learning low inference complexity Bayesian networks. For that, we use a representation of the network factorization that allows efficiently evaluating an upper bound in the inference complexity of each model during the learning process. Experimental results show that the proposed methods obtain tractable models that improve the accuracy of the predictions provided by approximate inference in models obtained with a well-known Bayesian network learner.
ISSN:2192-6352
2192-6360
DOI:10.1007/s13748-015-0070-0