Inductive Generalization of Analytically Learned Goal Hierarchies

  • Tolga Könik
  • Negin Nejati
  • Ugur Kuter
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5989)

Abstract

We describe a new approach for learning procedural knowledge represented as teleoreactive logic programs using relational behavior traces as input. This representation organizes task decomposition skills hierarchically and associate explicitly defined goals with them. Our approach integrates analytical learning with inductive generalization in order to learn these skills. The analytical component predicts the goal dependencies in a successful solution and generates a teleoreactive logic program that can solve similar problems by determining the structure of the skill hierarchy and skill applicability conditions (preconditions), which may be overgeneral. The inductive component experiments with these skills on new problems and uses the data collected in this process to refine the preconditions. Our system achieves this by converting the data collected during the problem solving experiments into the positive and negative examples of preconditions that can be learned with a standard Inductive Logic Programming system. We show that this conversion uses one of the main commitments of teleoreactive logic programs: associating all skills with explicitly defined goals. We claim that our approach uses less expert effort compared to a purely inductive approach and performs better compared to a purely analytical approach.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Choi, D., Langley, P.: Learning teleoreactive logic programs from problem solving. In: ILP 2005. LNCS (LNAI), vol. 3625, pp. 51–68. Springer, Heidelberg (2005)Google Scholar
  2. 2.
    Sammut, C., Hurst, S., Kedzier, D., Michie, D.: Learning to fly. In: Sleeman, D., Edwards, P. (eds.) Proceedings of the 9th International Conference on Machine Learning, pp. 385–393 (1992)Google Scholar
  3. 3.
    Isaac, A., Sammut, C.: Goal-directed learning to fly. In: Fawcett, T., Mishra, N. (eds.) Proceedings of the 20th International Conference on Machine Learning (2003)Google Scholar
  4. 4.
    Könik, T., Laird, J.E.: Learning goal hierarchies from structured observations and expert annotations. Machine Learning 64, 263–287 (2006)MATHCrossRefGoogle Scholar
  5. 5.
    Nejati, N., Langley, P., Könik, T.: Learning hierarchical task networks by observation. In: Proceedings of the Twenty-Third International Conference on Machine Learning (2006)Google Scholar
  6. 6.
    Genesereth, M.R., Love, N., Pell, B.: General game playing overview of the aaai competition. AI Magazine 26, 62–72 (2005)Google Scholar
  7. 7.
    Erol, K., Hendler, J., Nau, D.S.: HTN planning: Complexity and expressivity. In: Procceedings of the Nineteenth National Conference on Artificial Intelligence (1994)Google Scholar
  8. 8.
    Nau, D., Au, T.C., Ilghami, O., Kuter, U., Murdock, W., Wu, D., Yaman, F.: SHOP2: An HTN planning system. Journal of Artificial Intelligence Research 20, 379–404 (2003)MATHGoogle Scholar
  9. 9.
    Waisbrot, N., Kuter, U., Könik, T.: Combining heuristic search with hierarchical task-network planning: A preliminary report. In: Proceedings of the 21st International Florida Artificial Intelligence Research Society Conference (2008)Google Scholar
  10. 10.
    Nejati, N., Könik, T., Kuter, U.: A goal and dependency-driven algorithm for learning hierarchical task networks. In: Procceedings of the Fifth International Conference on Knowledge Capture, pp. 113–120. ACM, New York (2009)CrossRefGoogle Scholar
  11. 11.
    Dejong, G., Mooney, R.: Explanation-based learning: An alternative view. Machine Learning 1(2), 145–176 (1986)Google Scholar
  12. 12.
  13. 13.
    Sleeman, D., Langley, P., Mitchell, T.M.: Learning from solution paths: An approach to the credit assignment problem. AI Magazine 3(2) (1982)Google Scholar
  14. 14.
    Segre, A.M.: ARMS: Acquiring robotic assembly plans. In: De Jong, G. (ed.) Investigating Explanation-Based Learning. Kluwer, Boston (1993)Google Scholar
  15. 15.
    Mooney, R.J., Zelle, J.M.: Integrating ILP and EBL. SIGART Bulletin 5, 12–21 (1994)CrossRefGoogle Scholar
  16. 16.
    Estlin, T.A., Mooney, R.J.: Integrating EBL and ILP to acquire control rules for planning. In: Proceedings of the 3rd International Workshop on Multi-Strategy Learning, pp. 271–279 (1996)Google Scholar
  17. 17.
    Reddy, C., Tadepalli, P.: Learning goal-decomposition rules using exercises. In: Proceedings of the 14th International Conference on Machine Learning (1997)Google Scholar
  18. 18.
    Hogg, C., Muñoz Avila, H., Kuter, U.: HTN-MAKER: Learning HTNs with minimal additional knowledge engineering required. In: Proceedings of the 23rd AAAI Conference on Artificial Intelligence, pp. 950–956. AAAI Press, Menlo Park (2008)Google Scholar
  19. 19.
    Ilghami, O., Muñoz Avila, H., Nau, D., Aha, D.: Learning approximate preconditions for methods in hierarchical plans. In: Proceedings of the 22nd International Conference on Machine Learning, Bonn, Germany (2005)Google Scholar
  20. 20.
    Khardon, R.: Learning action strategies for planning domains. Artificial Intelligence 113, 125–148 (1999)MATHCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Tolga Könik
    • 1
  • Negin Nejati
    • 1
  • Ugur Kuter
    • 2
  1. 1.Computational Learning LaboratoryStanford UniversityStanford
  2. 2.Department of Computer Science and Institute for Advanced Computer StudiesUniversity of MarylandCollege Park

Personalised recommendations