Parameter Learning Algorithms for Continuous Model Improvement Using Operational Data

  • Anders L. Madsen
  • Nicolaj Søndberg Jeppesen
  • Frank Jensen
  • Mohamed S. Sayed
  • Ulrich Moser
  • Luis Neto
  • Joao Reis
  • Niels Lohse
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10369)


In this paper, we consider the application of object-oriented Bayesian networks to failure diagnostics in manufacturing systems and continuous model improvement based on operational data. The analysis is based on an object-oriented Bayesian network developed for failure diagnostics of a one-dimensional pick-and-place industrial robot developed by IEF-Werner GmbH. We consider four learning algorithms (batch Expectation-Maximization (EM), incremental EM, Online EM and fractional updating) for parameter updating in the object-oriented Bayesian network using a real operational dataset. Also, we evaluate the performance of the considered algorithms on a dataset generated from the model to determine which algorithm is best suited for recovering the underlying generating distribution. The object-oriented Bayesian network has been integrated into both the control software of the robot as well as into a software architecture that supports diagnostic and prognostic capabilities of devices in manufacturing systems. We evaluate the time performance of the architecture to determine the feasibility of on-line learning from operational data using each of the four algorithms.


Bayesian networks Parameter update Practical application 


  1. 1.
    Binder, J., Koller, D., Russell, S., Kanazawa, K.: Adaptive probabilistic networks with hidden variables. Mach. Learn. 29(2), 213–244 (1997)CrossRefMATHGoogle Scholar
  2. 2.
    Cappe, O., Moulines, E.: Online EM algorithm for latent data models. J. Roy. Stat. Soc. Ser. B (Stat. Method.) 71(3), 593–613 (2009)CrossRefMATHGoogle Scholar
  3. 3.
    Cowell, R.G., Dawid, A.P., Lauritzen, S.L., Spiegelhalter, D.J.: Probabilistic Networks and Expert Systems. Springer, New York (1999)MATHGoogle Scholar
  4. 4.
    Dempster, A.P., Laird, N.M., Rubin, D.B.: Maximum likelihood from incomplete data via the EM algorithm. J. Roy. Stat. Soc. Ser. B 39(1), 1–38 (1977)MathSciNetMATHGoogle Scholar
  5. 5.
    Jensen, F.V.: Gradient descent training of Bayesian networks. In: Proceedings of the ECSQARU, pp. 190–200 (1999)Google Scholar
  6. 6.
    Kjærulff, U.B., Madsen, A.L.: Bayesian Networks and Influence Diagrams. A Guide to Construction and Analysis, 2nd edn. Springer, New York (2013)MATHGoogle Scholar
  7. 7.
    Kokolakis, G., Nanopoulos, P.: Bayesian multivariate micro-aggregation under the Hellingers distance criterion. Res. Offic. Stat. 4(1), 117–126 (2001)Google Scholar
  8. 8.
    Koller, D., Pfeffer, A.: Object-oriented bayesian networks. In: Proceedings of the UAI, pp. 302–313 (1997)Google Scholar
  9. 9.
    Lauritzen, S.L.: The EM algorithm for graphical association models with missing data. Comput. Stat. Anal. 19, 191–201 (1995)CrossRefMATHGoogle Scholar
  10. 10.
    Madsen, A.L., Lang, M., Kjærulff, U.B., Jensen, F.: The Hugin tool for learning Bayesian networks. In: Proceedings of the ECSQARU, pp. 594–605 (2003)Google Scholar
  11. 11.
    Madsen, A.L., Søndberg-Jeppesen, N., Lohse, N., Sayed, M.: A methodology for developing local smart diagnostic models using expert knowledge. In: IEEE INDIN, pp. 1682–1687 (2015)Google Scholar
  12. 12.
    Madsen, A.L., Søndberg-Jeppesen, N., Sayed, M.S., Peschl, M., Lohse, N.: Applying object-oriented Bayesian networks for smart diagnosis and health monitoring at both component and factory level. Accepted for IEA/AIE 2017 (2017)Google Scholar
  13. 13.
    Neil, M., Fenton, N., Nielsen, L.M.: Building large-scale Bayesian networks. Knowl. Eng. Rev. 15(3), 257–284 (2000)CrossRefMATHGoogle Scholar
  14. 14.
    Neto, L., Reis, J., Guimaraes, D., Concalves, G.: Sensor cloud: smartcomponent framework for reconfigurable diagnostics in intelligent manufacturing environments. In: IEEE INDIN, pp. 1706–1711 (2015)Google Scholar
  15. 15.
    Neto, L., Reis, J., Silva, R., Concalves, G.: Sensor SelComp, a smart component for the industrial sensor cloud of the future. In: IEEE ICIT, pp. 1256–1261 (2017)Google Scholar
  16. 16.
    Olesen, K.G., Lauritzen, S.L., Jensen, F.V.: aHUGIN: a system creating adaptive causal probabilistic networks. In: Proceedings of the UAI, pp. 223–229 (1992)Google Scholar
  17. 17.
    Pearl, J.: Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufmann Publishers, San Mateo (1988)MATHGoogle Scholar
  18. 18.
    Ratnapinda, P., Druzdzel, M.J.: Learning discrete Bayesian network parameters from continuous data streams: what is the best strategy. J. Appl. Logic 13, 628–642 (2015)MathSciNetCrossRefMATHGoogle Scholar
  19. 19.
    Russell, S., Binder, J., Koller, D., Kanazawa, K.: Local learning in probabilistic networks with hidden variables. In: Proceedings of IJCAI, pp. 1146–1152 (1995)Google Scholar
  20. 20.
    Sayed, M.S., Lohse, N., Søndberg-Jeppesen, N., Madsen, A.L.: SelSus: towards a reference architecture for diagnostics and predictive maintenance using smart manufacturing devices. In: IEEE INDIN, p. 6 (2015)Google Scholar
  21. 21.
    Titterington, D.M.: Updating a diagnostic system using unconfirmed cases. Appl. Stat. 25, 238–247 (1976)CrossRefGoogle Scholar
  22. 22.
    Zagorecki, A., Voortman, M., Druzdzel, M.J.: Decomposing local probability distributions in bayesian networks for improved inference and parameter learning. In: Proceedings of the FLAIRS, pp. 860–865 (2006)Google Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Anders L. Madsen
    • 1
    • 2
  • Nicolaj Søndberg Jeppesen
    • 1
  • Frank Jensen
    • 1
  • Mohamed S. Sayed
    • 3
  • Ulrich Moser
    • 4
  • Luis Neto
    • 5
  • Joao Reis
    • 5
  • Niels Lohse
    • 3
  1. 1.HUGIN EXPERT A/SAalborgDenmark
  2. 2.Department of Computer ScienceAalborg UniversityAalborgDenmark
  3. 3.Loughborough UniversityLoughboroughUK
  4. 4.IEF-Werner GmbHFurtwangenGermany
  5. 5.Instituto de Sistemas e. Robotica AssociacaoPortoPortugal

Personalised recommendations