Abductive inference in Bayesian networks using distributed overlapping swarm intelligence

Fortier, Nathan; Sheppard, John W.; Strasser, Shane

doi:10.1007/s00500-014-1310-0

Cited by 18 publications

(13 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These methods include work like learning Bayesian network structures using Ant Colony Optimization [17] and abductive inference using PSO [8]. However, to our knowledge, no work has been published using PSO or multi-population methods for parameter estimation.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Parameter Estimation in Bayesian Networks Using Overlapping Swarm Intelligence

Fortier

Sheppard

Strasser

2015

Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

Bayesian networks are probabilistic graphical models that have proven to be able to handle uncertainty in many realworld applications. One key issue in learning Bayesian networks is parameter estimation, i.e., learning the local conditional distributions of each variable in the model. While parameter estimation can be performed efficiently when complete training data is available (i.e., when all variables have been observed), learning the local distributions becomes difficult when latent (hidden) variables are introduced. While Expectation Maximization (EM) is commonly used to perform parameter estimation in the context of latent variables, EM is a local optimization method that often converges to sub-optimal estimates. Although several authors have improved upon traditional EM, few have applied population based search techniques to parameter estimation, and most existing population-based approaches fail to exploit the conditional independence properties of the networks. We introduce two new methods for parameter estimation in Bayesian networks based on particle swarm optimization (PSO). The first is a single swarm PSO, while the second is a multiswarm PSO algorithm. In the multi-swarm version, a swarm is assigned to the Markov blanket of each variable to be estimated, and competition is held between overlapping swarms. Results of comparing these new methods to several existing approaches indicate that the multi-swarm algorithm outperforms the competing approaches when compared using data generated from a variety of Bayesian networks.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Following the success of OSI on training deep neural networks, Fortier et al adapted the OSI to yield a method for both partial and full abductive inference in Bayesian Networks [8]. In this approach, multiple swarms are used to find the most probable state assignments for a Bayesian network given the evidence.…”

Section: Multi-population Algorithmsmentioning

confidence: 99%

Parameter Estimation in Bayesian Networks Using Overlapping Swarm Intelligence

Fortier

Sheppard

Strasser

2015

Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

show abstract

“…Fortier and Sheppard developed a method for abductive inference in Bayesian Networks based on OSI [31]. In this approach, multiple swarms are used to find the most probable state assignments for a Bayesian network given the evidence.…”

Section: B Multi-population Algorithmsmentioning

confidence: 99%

Learning Bayesian classifiers using overlapping swarm intelligence

Fortier

Sheppard

Strasser

2014

2014 IEEE Symposium on Swarm Intelligence

Self Cite

View full text Add to dashboard Cite

Bayesian networks are powerful probabilistic models that have been applied to a variety of tasks. When applied to classification problems, Bayesian networks have shown competitive performance when compared to other state-of-theart classifiers. However, structure learning of Bayesian networks has been shown to be NP-Hard. In this paper, we propose a novel approximation algorithm for learning Bayesian network classifiers based on Overlapping Swarm Intelligence. In our approach a swarm is associated with each attribute in the data. Each swarm learns the edges for its associated attribute node and swarms that learn conflicting structures compete for inclusion in the final network structure. Our results indicate that, in many cases, Overlapping Swarm Intelligence significantly outperforms competing approaches, including traditional particle swarm optimization.

show abstract

“…Pseudocode for the Algorithm Peirce Figure 1 presents the pseudocode for the algorithm Peirce. The algorithm Peirce formulates hypotheses that comply with equations (5), (6), (7) and the criterion to select good hypotheses of the Definition 6. Synthetically, the algorithm Peirce formulates candidate hypotheses and stores them in set H (line 11).…”

Section: 3mentioning

confidence: 99%

“…Different approaches have been used to develop algorithms for abductive reasoning. Among the many contributions, there are proposals that use search techniques [15] and probabilistic reasoning over Bayesian Networks [7]. Logic approaches are based on two types of contributions: (1) proposal of new algorithms and (2) extension of traditional logical programming to process abductive reasoning problems.…”

Section: Related Workmentioning

confidence: 99%

Peirce: an Algorithm for Abductive Reasoning Operating with a Quaternary Reasoning Framework

Rodrigues¹,

Oliveira²,

Oliveira³

2014

RCS

View full text Add to dashboard Cite

Abductive reasoning algorithms formulate possible hypotheses to explain observed facts using a theory as the basis. These algorithms have been applied to various domains such as diagnosis, planning and interpretation. In general, algorithms for abductive reasoning based on logic present the following disadvantages: (1) they do not allow the explicit declaration of conditions that may affect the reasoning, such as intention, context and belief;(2) they allow little or no consideration for criteria required to select good hypotheses. Using Propositional Logic as its foundation, this study proposes the algorithm Peirce, which operates with a framework that allows one to explicitly include conditions to conduct abductive reasoning and uses a criterion to select good hypotheses that employs metrics to define the explanatory power and complexity of the hypotheses. Experimental results suggest that abductive reasoning performed by humans has the tendency to coincide with the solutions computed by the algorithm Peirce.

show abstract

Abductive inference in Bayesian networks using distributed overlapping swarm intelligence

Cited by 18 publications

References 25 publications

Parameter Estimation in Bayesian Networks Using Overlapping Swarm Intelligence

Parameter Estimation in Bayesian Networks Using Overlapping Swarm Intelligence

Learning Bayesian classifiers using overlapping swarm intelligence

Peirce: an Algorithm for Abductive Reasoning Operating with a Quaternary Reasoning Framework

Contact Info

Product

Resources

About