Bayesian Network Languages: Inference and Complexity


  • Bayesian Network Languages: Inference and Complexity

Bayesian Network Languages: Inference and Complexity

Principal Supervisor: Dr. Cassio P. de Campos              

Second Supervisor: Dr.Hans Vandierendonck

+ Project Description

A Bayesian network is a probabilistic model with applications in a wide range of domains. The main features of Bayesian networks are compact representation even for large multi-variate domains, possibility of unveiling dependences and independences among variables, fast algorithms based on graphs. Bayesian networks have two components: a graph and probability tables. We can contemplate many languages to specify conditional probabilities

that go beyond flat conditional probability tables. In this project we wish to study the properties of inference algorithms

as they are parameterized by such specification languages. In particular, we wish to study the relationship between specification languages and complexity. By allowing new constructs we are extending the Bayesian network applicability to even further domains. By restricting existing constructs, we may gain performance and save computational resources. By studying computational complexity under such extensions/restrictions, we will be able to identify easy and hard problems, thus creating the foundations to design efficient inference algorithms. We will dedicate attention to languages that can be implemented efficiently using current technology of compilers and hardware. 

+ How to Apply

Applicants should apply electronically through the Queen’s online application portal at:

+ Contact Details

Supervisor Name: Dr. Cassio P de Campos

Queens University of Belfast
School of EEECS
18 Malone Road, CSB 03.026           


+44 (0)28 9097 6795