Comparison of Machine Learning for Autonomous Robot Discovery

  • Ivan Bratko
Part of the Studies in Computational Intelligence book series (SCI, volume 262)


In this paper we consider autonomous robot discovery through experimentation in the robot’s environment. We analyse the applicability of machine learning (ML) methods with respect to various levels of robot discovery tasks, from extracting simple laws among the observed variables, to discovering completely new notions that were never explicitly mentioned in the data directly. We first present some illustrative experiments in robot learning in the XPERO European project. Then we formulate criteria for a comparison of learning methods and a systematic list of types of learning or discovery tasks, and discuss the suitability of chosen ML methods for these tasks.


Machine learning robotic discovery autonomous learning gaining insights 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Awaad, I., Leon, B.: XPERSim: Simulation of the robotic experimenter. XPERO Report on R&D 1 (September 2006)Google Scholar
  2. 2.
    Bratko, I.: Prolog Programming for Artificial Intelligence, 3rd edn. Addison-Wesley / Pearson, Reading (2001)Google Scholar
  3. 3.
    Bratko, I., Šuc, D.: Learning qualitative models. AI Magazine 24(4), 107–119 (2003)Google Scholar
  4. 4.
    Bratko, I., Šuc, D., Awaad, I., Demšar, J., Gemeiner, P., Guid, M., Leon, B., Mestnik, M., Prankl, J., Prassler, E., Vincze, M., Žabkar, J.: Initial experiments in robot discovery in XPERO. In: ICRA 2007 workshop Concept Learning for Embodied Agents, Rome (April 2007)Google Scholar
  5. 5.
    Clark, P., Niblett, T.: The CN2 induction algorithm. Machine Learning 3, 262–284 (1989)Google Scholar
  6. 6.
    Demšar, J., Zupan, B.: Orange: Data Mining Fruitful & Fun - From Experimental Machine Learning to Interactive Data Mining (2006),
  7. 7.
    Džeroski, S., Todorovski, L.: Discovering dynamics. In: Proceedings of the 10th International Conference on Machine Learning, pp. 97–103. Morgan Kaufmann, San Francisco (1993)Google Scholar
  8. 8.
    Križman, V., Džeroski, S., Kompare, B.: Discovering dynamics from measured data. Electrotechnical Review 62, 191–198 (1995)Google Scholar
  9. 9.
    Križman, V.: Automatic Discovery of the Structure of Dynamic System Models. PhD thesis, Faculty of Computer and Information Sciences, University of Ljubljana (1998)Google Scholar
  10. 10.
    Kuipers, B.: Qualitative Reasoning: Modeling and Simulation with Incomplete Knowledge. MIT Press, Cambridge (1994)Google Scholar
  11. 11.
    Leban, G., Žabkar, J., Bratko, I.: An experiment in robot discovery with ILP. In: Železný, F., Lavrač, N. (eds.) ILP 2008. LNCS (LNAI), vol. 5194, pp. 77–90. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  12. 12.
    Muggleton, S.: Inverse entailment and Progol. New Generation Computing 13, 245–286 (1995)CrossRefGoogle Scholar
  13. 13.
    Srinivasan, A.: The Aleph Manual. Technical Report, Computing Laboratory, Oxford University (2000),
  14. 14.
    Šuc, D.: Machine Reconstruction of Human Control Strategies, Ph.D. Thesis, Faculty of Computer and Information Sc., University of Ljubljana (2001)Google Scholar
  15. 15.
    Šuc, D.: Machine Reconstruction of Human Control Strategies. Frontiers Artificial Intelligence Appl., vol. 99. IOS Press, Amsterdam (2003)Google Scholar
  16. 16.
    Šuc, D., Vladušič, D., Bratko, I.: Qualitatively Faithful Quantitative Prediction. Artificial Intelligence 158(2), 190–219 (2004)Google Scholar
  17. 17.
    Žabkar, J., Bratko, I., Demšar, J.: Learning qualitative models through partial derivatives by Padé. In: Proc. QR 2007, 21st International Workshop on Qualitative Reasoning, Aberystwyth, Wales, UK, pp. 193–202 (2007)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Ivan Bratko
    • 1
  1. 1.Faculty of Computer and Information Sc.University of LjubljanaLjubljanaSlovenia

Personalised recommendations