Signal Processing and Speech Communication Laboratory
homeresearch projects › Discriminative Learning of Bayesian Network Classifiers

Discriminative Learning of Bayesian Network Classifiers

Period
2007 — 2010
Funding
FWF Austrian Science Fund, Project P19737-N15
Research Areas
Contact

Over the last decade, Bayesian networks have become the method of choice for representation of uncertainty in machine learning. Bayesian networks are used in many research areas such as bioinformatics, computer vision, speech recognition, error-correcting coding theory, and artificial intelligence. Currently, the research is focused on two main issues. First, much work is devoted to finding more efficient approximate inference algorithms. Second, there has been much interest in learning the parameters and the structure of Bayesian networks from data. Basically, there are two main paradigms for learning in the machine learning community: generative and discriminative learning. There is a strong belief in the scientific community that discriminative classifiers have to be preferred in reasoning tasks. The aim of the proposed research is to work on discriminative structure and parameter learning methods for Bayesian networks and to propose conditions for discriminative structures to be sufficient even trained only with maximum likelihood parameter training. Additionally, we want to perform an extensive experimental comparison between the developed discriminative approaches and well known generative methods. For the experiments, we want to use data sets from the UCI repository and from a surface inspection task available at our institute.

Related publications
  • Article Pernkopf F. & Bilmes J. (2010) Efficient Heuristics for Discriminative Structure Learning of Bayesian Network Classifiers. in Journal of machine learning research, p. 2323-2360. [more info]
  • Conference contribution Pernkopf F. & Wohlmayr M. (2009) On Discriminative Parameter Learning of Bayesian Network Classifiers. in European Conference on Machine Learning (ECML 2009) (pp. 221-237). [more info]
  • Conference contribution Wohlmayr M. & Pernkopf F. (2009) Finite Mixture Spectrogram Modeling for Multipitch Tracking Using A Factorial Hidden Markov Model. in Proceedings of the 10th Annual Conference of the International Speech Communication Association (pp. 1079-1082). [more info]
  • Conference contribution Pernkopf F. & Bilmes J. (2008) Order-based Discriminative Structure Learning for Bayesian network Classifiers. in International Symposium on Artificial Intelligence and Mathematics (pp. 1-8). [more info]
  • Conference contribution Wohlmayr M. & Pernkopf F. (2008) Multipitch Tracking Using A Factorial Hidden Markov Model. in Interspeech - International Conference on Spoken Language Processing (pp. 1-4). [more info]
  • Article Pernkopf F., Pham T. & Bilmes J. (2008) Broad Phonetic Classification Using Discriminative Bayesian Networks. in Speech Communication, 51(2), p. 151-166. [more info]
  • Conference contribution Pernkopf F. & Bilmes J. (2005) Discriminative versus Generative Parameter and Structure Learning of Bayesian Network Classifiers. in International Conference on Machine Learning (ICML) (pp. 657-664). [more info]