Preface: Special Issue on Advanced Methodologies for Bayesian Networks
Maomi Ueno 0 1 2
0 University of Electro-Communications , 1-5-1, Chofugaoka, Chofu-shi, Tokyo 182-8585 , Japan
1 Over the last few decades, Bayesian networks (BNs) have become an increasingly popular AI approach for treating uncertainty around random variables. The International Workshop on Advanced Methodologies for Bayesian Networks (AMBN 2010) was held on November 18-19, 2010, at the Campus Innovation Center in Tokyo , Japan [1, 2]. In AMBN 2010 , we concentrated on exploring methodologies for enhancing the effectiveness of BNs , including modeling, reasoning, model selection, logic- probability relations, and causality
2 - Changhe Yuan (Queens College/CUNY, USA) - Jiji Zhang (Lingnan University, Hong Kong) - Kun Zhang (University of Southern California , USA)
The second AMBN was held in Yokohama, Japan on November 16-18, 2015, co-sponsored by the Japanese Society Artificial Intelligence (JSAI) and the National Institute of Advanced Industrial Science and Technology (AIST).The AMBN workshop had over 100 participants and many lively and interesting discussions over the course of its duration. This special issue of New Generation Computing (NGC) has come out from these extensive activities of the researchers in BNs field.After the workshop, six excellent papers in the conference were recommended to submit to this special issue with the requirement at least 30% extension from the Post Proceeding of AMBN2015 (LNAI, Springer) version. Each paper was reviewed by one meta-reviewer and two reviewers. As a result, all the six papers were accepted for publication in this journal.
Artificial intelligence model; Bayesian networks; Machine learning; Probabilistic graphical
The second AMBN was held in Yokohama, Japan on November 16–18, 2015,
cosponsored by the Japanese Society Artificial Intelligence (JSAI) and the National
Institute of Advanced Industrial Science and Technology (AIST) [3, 4].
In AMBN 2015, in addition to exploring advanced methodologies of BNs, we
discussed practical considerations for applying BNs in real-world settings, covering
concerns such as scalability, incremental learning, and parallelization.
In spite of the short announcement period, 29 papers were submitted, and only 15
papers were accepted in the AMBN proceedings, which was published in the
Lecture Notes in Artificial Intelligence series by Springer. Each submission
underwent rigorous by three members of the AMBN Program Committee, with each
PC member reviewing at most two papers.
The members of Program Committee (PC) are leading researchers in BNs field:
The AMBN workshop had over 100 participants and many lively and interesting
discussions over the course of its duration. This special issue of New Generation
Computing (NGC) has come out from these extensive activities of the researchers in
After the workshop, six excellent papers in the conference were recommended to
submit to this special issue with the requirement at least 30% extension from the
Post Proceeding of AMBN2015 (LNAI, Springer) version.
Each paper was reviewed by one meta-reviewer and two reviewers. As a result,
all the six papers were accepted for publication in this journal.
The first paper titled as ‘‘On Model Selection, Bayesian Networks, and the Fisher
Information Integral’’ provides a very fundamental analysis of model selection for
Bayesian networks from information theoretical approach. This paper addresses
Bayesian Information Criterion (BIC) model selection that includes a constant term
involving the Fisher information matrix. They find that, for complex Bayesian
network models, the constant term is a negative number with a very large absolute
value that dominates the other terms for small and moderate sample sizes. For
networks with a fixed number of parameters, their experiments show that the
constant term can vary significantly depending on the network structure. In
particular, star-like networks have smaller complexity than networks where the node
degree is more uniform.
The second paper titled as ‘‘Learning Causal Graphs with Latent Confounder
Information in Faithfulness Violations’’ addressed a fundamental and important
analyses for causal graphs with latent common causes. Ancestral graph models are
effective and useful for representing causal models with some information of such
latent variables. The causal faithfulness condition, which is usually assumed for
determining the models, is known to often be weakly violated in statistical view
points for finite data. One of the authors developed a constraint-based causal
learning algorithm that is robust against the weak violations while assuming no
latent variables. In this study, we applied and extended the thoughts of the algorithm
to the inference of ancestral graph models. The practical validity and effectiveness
of the algorithm are also confirmed using some standard datasets in comparison with
the other traditional methods.
The third paper titled as ‘‘Duplicate Detection for Bayesian Network Structure
Learning’’ presents a new duplicate detection technique for Breadth-first branch and
bound in score-based Bayesian network structure learning problem. Previously, an
external sorting-based technique was used for delayed duplicate detection (DDD).
They propose a hashing-based technique for DDD and a bin packing algorithm for
minimizing the number of external memory files and operations. They also give a
structured duplicate detection approach which completely eliminates DDD.
Empirically, they demonstrate that structured duplicate detection is significantly
faster than the previous state of the art in limited-memory settings. Their results
show that the bin packing algorithm incurs some overhead, but that the overhead is
offset by reducing I/O when more memory is available.
The fourth paper titled as ‘‘Joint Analysis of Multiple Algorithms and
Performance Measures’’ develops statistical procedures that are able to account
for multiple competing measures at the same time and to compare multiple
algorithms altogether. In particular, they propose two tests: a frequentist procedure
based on the generalized likelihood-ratio test and a Bayesian procedure based on a
multinomial-Dirichlet conjugate model. They further extend them by discovering
conditional independences among measures to reduce the number of parameters of
such models, as usually the number of studied cases is very reduced in such
comparisons. Real data from a comparison among general purpose classifiers is used
to show a practical application of their tests.
The fifth paper titled ‘‘Improving Record Linkage Accuracy with Hierarchical
Feature Level Information and Parsed Data’’ addresses probabilistic record linkage
with hierarchical feature level information. This study extends the naive Bayes
classifier with such hierarchical feature level information. Moreover, they illustrate
the benefits of their method over previously proposed methods on 4 datasets in
terms of the linkage performance. The results show an improved performance of the
methods considered on further parsed datasets
The sixth paper titled ‘‘Efficient Bayesian Network Structure Learning for
Maximizing the Posterior Probability’’ addresses the problem of efficiently finding
an optimal Bayesian network structure for maximizing the posterior probability
using a Branch and Bound (B & B) technique. To make the search more efficient, B
& B technique needs a tighter upper bound so that the current score can exceed it
more easily. This paper proposes two upper bounds and prove that they are tighter
than the existing one. Moreover, this paper demonstrate that the proposed two
bounds render the search to be much more efficient using the Alarm and other major
data sets. For example, the search is three to four times faster for n ¼ 100 and two to
three times faster for n ¼ 500.
We would like to thank all of the authors and the reviewers of the published
papers. We also thank for the support of the Editorial board of NGC, the Editorial
Maomi Ueno, Guest Editor
1. Ueno , M. , Isozaki , T.: First International Workshop on Advanced Methodologies for Bayesian Networks. JSAI-isAI Workshops , vol. 2010 , pp. 165 - 166 ( 2010 )
2. Ueno , M. : Special issue on advanced methodologies for Bayesian networks . New Gener. Comput . 30 ( 1 ), 1 - 2 ( 2012 )
4. Suzuki , J. , Ueno , M. : Advanced Methodologies for Bayesian networks . In: Second International Workshop, AMBN2015 Proceedings, LNAI 9505 . Springer ( 2015 )