PGM 2008

The Fourth European Workshop on Probabilistic Graphical Models

Preliminary Program

Wednesday 17/9


9-10.30  Learning



Large incomplete sample robustness in Bayesian networks
Jim Smith1 and Alireza Daneshkhah2
1University of Warwick, Uk, 2Strathclyde University, UK


Factorized Normalized Maximum Likelihood Criterion for Learning Bayesian Network Structures
Tomi Silander,  Teemu Roos,  Petri Kontkanen,  Petri Myllymäki
University of Helsinki


Logical Properties of Stable Conditional Independence
Mathias Niepert and Dirk Van Gucht
Indiana University


11.00-12.30  Learning



Structural-EM for Learning PDG Models from Incomplete Data
Jens Nielsen1,  Rafael Rumí2,  Antonio Salmerón2
1Universidad de Castilla-La Mancha, 2Universidad de Almería


A Geometric Approach to Learning BN Structures
Milan Studeny and Jiri Vomlel
Institute of Information Theory and Automation of the ASCR


Robust Classification Using Mixtures of Dependency Networks
Jose A. Gamez1,   Juan L. Mateo1,   Thomas D. Nielsen2,   Jose M. Puerta1
1University of Castilla-La Mancha, 2Aalborg University


15.00-16.00  Inference



Policy Explanation in Factored Markov Decision Processes
Francisco Elizalde1,  Enrique Sucar2,  Manuel Luque3,  Javier Díez3,  Alberto Reyes1
1IIE, 2INAOE, 3UNED


Approximate representation of optimal strategies from influence diagrams
Finn Verner Jensen
Aalborg University


16.30-17.30  Parameter estimation



Parameter Estimation in Mixtures of Truncated Exponentials
Helge Langseth1, Thomas D. Nielsen2, Rafael Rumí3, Antonio Salmeron3
1The Norwegian University of Science and Technology, 2Aalborg University, 3University of Almeria


Discrimination and its sensitivity in probabilistic networks
Silja Renooij and Linda C. van der Gaag
Utrecht University





Thursday 18/9


9.00-10.30  Multi-agent systems



An Influence Diagram Framework for acting under influence by agents with unknown goals
Nicolaj Søndberg-Jeppesen and Finn Verner Jensen
Aalborg University


Tightly and Loosely Coupled Decision Paradigms in Multiagent Expedition
Yang Xiang and Franklin Hanshar
University of Guelph


Spotlight presentations


11.00-12.30  Modeling



Bayesian Networks: the Quantitative Parental Synergy
Janneke H. Bolt
Utrecht University


Arithmetic circuits of the noisy-or models
Jirka Vomlel and Petr Savicky
Academy of Sciences of the Czech Republic


Spotlight presentations


16.00-19.00  Poster session



An anytime algorithm for evaluating unconstrained influence diagrams
Manuel Luque1, Thomas D. Nielsen2, Finn V. Jensen2
1Departmento Inteligencia Artificial, UNED, 2Aalborg University


Solving CLQG Influence Diagrams Using Arc-Reversal
Anders Madsen
HUGIN EXPERT


Perturbing a set of inaccuracies parameters in Gaussian Bayesian networks
Rosario Susi,  Miguel A. Gómez-Villegas,  Paloma Main
Universidad Complutense de Madrid


The Probabilistic Interpretation of Model-based Diagnosis
Ildiko Flesch1 and Peter Lucas2
1University of Maastricht, 2Radboud University Nijmegen


Towards consistency in general dependency networks
José A. Gámez,  Juan L. Mateo,  José M. Puerta
University of Castilla-La Mancha


A* Wars: The Fight for Improving A* Search for Troubleshooting with Dependent Actions
Thorsten J. Ottosen and Finn V. Jensen
Department of Computer Science, Aalborg University, Denmark


Marginals of DAG-isomorphic Independence Models
Peter R. de Waal
Universiteit Utrecht


Complexity Results for Enumerating MPE and Partial MAP
Johan Kwisthout
Utrecht University


Computing the Multinomial Stochastic Complexity in Sub-Linear Time
Tommi Mononen and Petri Myllymäki
Helsinki Institute for Information Technology


Eliciting expert beliefs on the structure of a Bayesian Network
Federico M. Stefanini
Department of Statistics "G.Parenti", University of Florence


Efficient Bayesian Network Learning Using Pairwise Deletion
Olivier Francois
University of Reading


Carmen: An open source project for probabilistic graphical models
Manuel Arias and Francisco Javier Diez
UNED


High-dimensional probability density estimation with randomized ensembles of tree structured Bayesian networks
Sourour Ammar1,  Philippe Leray1,  Boris Defourny2,  Louis Wehenkel2
1LINA, Univ. Nantes, France, 2DEECS & GIGA-R, Univ. Liège, Belgium


A Bayesian approach to estimate probabilities in classification trees
Andrés Cano,  Andrés Masegosa,  Serafín Moral
University of Granada


A novel scalable and correct Markov boundary learning algorithms under faithfulness condition
Sergio Rodrigues de Morais1 and Alex Aussem2
1INSA-Lyon, LIESP, 2Université de Lyon 1, LIESP


Attribute Clustering Based on Heuristic Tree Partition
Jorge Cordero Hernandez and Yifeng Zeng
Aalborg University


An ICI Model for Opposing Influences
Paul Maaskant1 and Marek Druzdzel2
1Delft University of Technology, 2University of Pittsburgh


Query-based Diagnostics
John Mark Agosta1,  Tom Gardos1,  Marek Druzdzel2
1Intel Research, 2University of Pittsburgh


Measuring Efficiency in Influence Diagram Models
Barry Cobb
Virginia Military Institute





Friday 19/9


9.00-10.30  Inference



An empirical analysis of loopy belief propagation in three topologies: grids, small-world networks and random graphs
Roberto Santana,  Alexander Mendiburu,  Jose A. Lozano
The University of the Basque Country


Generalized Loopy 2U: A New Algorithm for Approximate Inference in Credal Networks
Alessandro Antonucci,  Marco Zaffalon,  Yi Sun,  Cassio P. de Campos
IDSIA, Switzerland


New Methods for Marginalization in Lazy Propagation
Anders Madsen
HUGIN EXPERT


11.00-12.30  Learning



Learning naive Bayes regression models from missing data using mixtures of truncated exponentials
Antonio Fernández1,  Jens D. Nielsen2,  Antonio Salmerón1
1University of Almería, 2University of Castilla-La Mancha


Efficient Model Evaluation in the Search-Based Approach to Latent Structure Discovery
Tao Chen,  Nevin L. Zhang,  Yi Wang
HKUST


A Score Based Ranking of the Edges for the PC Algorithm
Andres Cano,  Manuel Gomez-Olmedo,  Serafín Moral
Universidad de Granada